docs: record Qwen3.5-9B DFlash Metal timeout (refs #152, #154)

Merged PR #90: feat: integration test — turboquant compressed model

.gitea/workflows/smoke.yml

@@ -22,7 +22,3 @@ jobs:
         run: |
           if grep -rE 'sk-or-|sk-ant-|ghp_|AKIA' . --include='*.yml' --include='*.py' --include='*.sh' 2>/dev/null | grep -v .gitea | grep -v llama-cpp-fork; then exit 1; fi
           echo "PASS: No secrets"
-      - name: Tool call regression (schema validation)
-        run: |
-          python3 tests/tool_call_regression.py --dry-run
-          echo "PASS: Tool call schemas valid"

README.md

@@ -30,3 +30,4 @@ See [issues](https://forge.alexanderwhitestone.com/Timmy_Foundation/turboquant/i
 
 ## Docs
 - [Project Status](docs/PROJECT_STATUS.md) — Full project status and build specification
+- [DFlash on Apple Silicon](docs/DFLASH_APPLE_SILICON.md) — MLX benchmark planner, setup commands, and report workflow

benchmarks/dflash_apple_silicon.py

@@ -0,0 +1,189 @@
+#!/usr/bin/env python3
+"""Apple Silicon DFlash planning helpers and CLI (issue #152)."""
+
+from __future__ import annotations
+
+import argparse
+import json
+import platform
+import subprocess
+from dataclasses import asdict, dataclass
+from pathlib import Path
+from typing import Iterable, Optional
+
+
+@dataclass(frozen=True)
+class DFlashPair:
+    slug: str
+    base_model: str
+    draft_model: str
+    estimated_total_weights_gb: float
+    minimum_recommended_memory_gb: float
+    draft_sliding_window_size: int = 4096
+
+
+SUPPORTED_PAIRS: tuple[DFlashPair, ...] = (
+    DFlashPair(
+        slug="qwen35-4b",
+        base_model="Qwen/Qwen3.5-4B",
+        draft_model="z-lab/Qwen3.5-4B-DFlash",
+        estimated_total_weights_gb=9.68,
+        minimum_recommended_memory_gb=16.0,
+    ),
+    DFlashPair(
+        slug="qwen35-9b",
+        base_model="Qwen/Qwen3.5-9B",
+        draft_model="z-lab/Qwen3.5-9B-DFlash",
+        estimated_total_weights_gb=19.93,
+        minimum_recommended_memory_gb=28.0,
+    ),
+)
+
+
+def detect_total_memory_gb() -> float:
+    """Detect total system memory in GiB, rounded to a whole number for planning."""
+    system = platform.system()
+    if system == "Darwin":
+        mem_bytes = int(subprocess.check_output(["sysctl", "-n", "hw.memsize"]).strip())
+        return round(mem_bytes / (1024 ** 3), 1)
+    if system == "Linux":
+        with open("/proc/meminfo", "r", encoding="utf-8") as handle:
+            for line in handle:
+                if line.startswith("MemTotal:"):
+                    mem_kb = int(line.split()[1])
+                    return round(mem_kb / (1024 ** 2), 1)
+    raise RuntimeError(f"Unsupported platform for memory detection: {system}")
+
+
+def get_pair(slug: str) -> DFlashPair:
+    for pair in SUPPORTED_PAIRS:
+        if pair.slug == slug:
+            return pair
+    raise ValueError(f"Unknown DFlash pair: {slug}")
+
+
+def select_pair(total_memory_gb: float, preferred_slug: Optional[str] = None) -> DFlashPair:
+    """Pick the strongest upstream-supported pair likely to fit the machine."""
+    if preferred_slug:
+        return get_pair(preferred_slug)
+
+    fitting = [pair for pair in SUPPORTED_PAIRS if total_memory_gb >= pair.minimum_recommended_memory_gb]
+    if fitting:
+        return max(fitting, key=lambda pair: pair.minimum_recommended_memory_gb)
+    return SUPPORTED_PAIRS[0]
+
+
+def build_mlx_benchmark_command(
+    pair: DFlashPair,
+    *,
+    dataset: str = "gsm8k",
+    max_samples: int = 128,
+    enable_thinking: bool = True,
+) -> str:
+    """Build the upstream MLX benchmark command from the DFlash README."""
+    parts = [
+        "python -m dflash.benchmark --backend mlx",
+        f"--model {pair.base_model}",
+        f"--draft-model {pair.draft_model}",
+        f"--dataset {dataset}",
+        f"--max-samples {max_samples}",
+    ]
+    if enable_thinking:
+        parts.append("--enable-thinking")
+    parts.append(f"--draft-sliding-window-size {pair.draft_sliding_window_size}")
+    return " \\\n    ".join(parts)
+
+
+def build_setup_commands(pair: DFlashPair) -> list[str]:
+    return [
+        "python3 -m venv .venv-dflash",
+        "source .venv-dflash/bin/activate",
+        "git clone https://github.com/z-lab/dflash.git",
+        "cd dflash",
+        "pip install -e .[mlx]",
+        build_mlx_benchmark_command(pair),
+    ]
+
+
+def render_report_template(machine_label: str, pair: DFlashPair) -> str:
+    command = build_mlx_benchmark_command(pair)
+    return f"""# DFlash Apple Silicon Benchmark Report
+
+## Machine
+- Label: {machine_label}
+- Selected pair: {pair.slug}
+- Base model: {pair.base_model}
+- Draft model: {pair.draft_model}
+- Estimated total weight footprint: {pair.estimated_total_weights_gb:.2f} GB
+
+## Setup
+```bash
+python3 -m venv .venv-dflash
+source .venv-dflash/bin/activate
+git clone https://github.com/z-lab/dflash.git
+cd dflash
+pip install -e .[mlx]
+{command}
+```
+
+## Baseline comparison
+Compare against **plain MLX or llama.cpp speculative decoding** on the same prompt set.
+
+## Results
+- Throughput (tok/s):
+- Peak memory (GB):
+- Notes on acceptance / behavior:
+
+## Verdict
+Worth operationalizing locally?
+- [ ] Yes
+- [ ] No
+- [ ] Needs more data
+
+## Recommendation
+Explain whether this should become part of the local inference stack.
+"""
+
+
+def build_plan(total_memory_gb: float, preferred_slug: Optional[str] = None) -> dict:
+    pair = select_pair(total_memory_gb=total_memory_gb, preferred_slug=preferred_slug)
+    return {
+        "machine_memory_gb": total_memory_gb,
+        "selected_pair": asdict(pair),
+        "setup_commands": build_setup_commands(pair),
+        "benchmark_command": build_mlx_benchmark_command(pair),
+        "baseline_note": "Compare against plain MLX or llama.cpp speculative decoding on the same prompt set.",
+    }
+
+
+def write_output(path: Path, content: str) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_text(content, encoding="utf-8")
+
+
+def main(argv: Optional[Iterable[str]] = None) -> int:
+    parser = argparse.ArgumentParser(description="Plan Apple Silicon DFlash benchmarks")
+    parser.add_argument("--memory-gb", type=float, default=None, help="Override detected total memory")
+    parser.add_argument("--pair", choices=[pair.slug for pair in SUPPORTED_PAIRS], default=None)
+    parser.add_argument("--machine-label", default="Apple Silicon Mac")
+    parser.add_argument("--format", choices=["json", "markdown"], default="markdown")
+    parser.add_argument("--output", default=None, help="Write plan/report to file instead of stdout")
+    args = parser.parse_args(list(argv) if argv is not None else None)
+
+    memory_gb = args.memory_gb if args.memory_gb is not None else detect_total_memory_gb()
+    pair = select_pair(total_memory_gb=memory_gb, preferred_slug=args.pair)
+
+    if args.format == "json":
+        content = json.dumps(build_plan(memory_gb, preferred_slug=pair.slug), indent=2)
+    else:
+        content = render_report_template(args.machine_label, pair)
+
+    if args.output:
+        write_output(Path(args.output), content)
+    else:
+        print(content)
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

benchmarks/reports/dflash_m3max_36gb.md

@@ -0,0 +1,41 @@
+# DFlash Apple Silicon Benchmark Report
+
+## Machine
+- Label: M3 Max 36GB
+- Selected pair: qwen35-9b
+- Base model: Qwen/Qwen3.5-9B
+- Draft model: z-lab/Qwen3.5-9B-DFlash
+- Estimated total weight footprint: 19.93 GB
+
+## Setup
+```bash
+python3 -m venv .venv-dflash
+source .venv-dflash/bin/activate
+git clone https://github.com/z-lab/dflash.git
+cd dflash
+pip install -e .[mlx]
+python -m dflash.benchmark --backend mlx \
+    --model Qwen/Qwen3.5-9B \
+    --draft-model z-lab/Qwen3.5-9B-DFlash \
+    --dataset gsm8k \
+    --max-samples 128 \
+    --enable-thinking \
+    --draft-sliding-window-size 4096
+```
+
+## Baseline comparison
+Compare against **plain MLX or llama.cpp speculative decoding** on the same prompt set.
+
+## Results
+- Throughput (tok/s):
+- Peak memory (GB):
+- Notes on acceptance / behavior:
+
+## Verdict
+Worth operationalizing locally?
+- [ ] Yes
+- [ ] No
+- [ ] Needs more data
+
+## Recommendation
+Explain whether this should become part of the local inference stack.

benchmarks/reports/dflash_m3max_36gb_qwen35_4b_pilot.md

@@ -0,0 +1,46 @@
+# DFlash Apple Silicon Pilot — Qwen3.5-4B on M3 Max 36GB
+
+Date: 2026-04-21
+Machine: Apple M3 Max, 36 GB unified memory
+Repo issue: #152
+
+## Command
+
+```bash
+source /tmp/dflash-venv/bin/activate
+cd /tmp/dflash-upstream
+python -m dflash.benchmark --backend mlx \
+    --model Qwen/Qwen3.5-4B \
+    --draft-model z-lab/Qwen3.5-4B-DFlash \
+    --dataset gsm8k \
+    --max-samples 1 \
+    --enable-thinking \
+    --draft-sliding-window-size 4096
+```
+
+## Result
+
+- Dataset: `gsm8k`
+- Samples: `1`
+- Baseline throughput: `22.35 tok/s`
+- DFlash throughput: `46.78 tok/s`
+- Decoding speedup: `2.09x`
+- Average acceptance length: `6.48`
+
+Acceptance length histogram:
+
+```text
+['0.3%', '11.1%', '12.7%', '10.4%', '11.7%', '7.6%', '7.0%', '3.8%', '5.1%', '6.3%', '2.8%', '3.8%', '2.2%', '1.9%', '0.9%', '2.5%', '9.8%']
+```
+
+## Caveats
+
+- This is a **pilot**, not a decision-grade benchmark.
+- Only `1` sample was run, so the throughput number is directional.
+- No apples-to-apples baseline against plain MLX or llama.cpp speculative decoding is included yet.
+- The planner still recommends trying `Qwen/Qwen3.5-9B + z-lab/Qwen3.5-9B-DFlash` on this machine for the more meaningful fit test.
+
+## Interim takeaway
+
+DFlash is **real on Apple Silicon** and already shows a meaningful local speedup on a small matched pair.
+A `2.09x` pilot speedup on `Qwen3.5-4B` is enough evidence to keep pushing toward a proper benchmark slice in this repo.

benchmarks/reports/dflash_m3max_36gb_qwen35_9b_timeout.md

@@ -0,0 +1,59 @@
+# DFlash on Apple Silicon Failure Report — Qwen3.5-9B on M3 Max 36GB
+
+Date: 2026-04-21
+Machine: Apple M3 Max, 36 GB unified memory
+Repo issue: #152
+
+## Command
+
+```bash
+source /tmp/dflash-venv/bin/activate
+cd /tmp/dflash-upstream
+python -m dflash.benchmark --backend mlx \
+    --model Qwen/Qwen3.5-9B \
+    --draft-model z-lab/Qwen3.5-9B-DFlash \
+    --dataset gsm8k \
+    --max-samples 1 \
+    --enable-thinking \
+    --draft-sliding-window-size 4096
+```
+
+## Outcome
+
+The benchmark did **not** complete successfully on this machine.
+
+### Failure signature
+
+```text
+libc++abi: terminating due to uncaught exception of type std::runtime_error:
+[METAL] Command buffer execution failed:
+Caused GPU Timeout Error (00000002:kIOGPUCommandBufferCallbackErrorTimeout)
+```
+
+Additional shutdown noise:
+
+```text
+bash: [11285: 1] tcsetattr: Inappropriate ioctl for device
+resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown
+```
+
+## Interpretation
+
+This is strong evidence that the `Qwen/Qwen3.5-9B + z-lab/Qwen3.5-9B-DFlash` pair is **not currently stable** on an M3 Max 36GB Mac under the upstream MLX benchmark path, at least with the default settings used here.
+
+It may still be salvageable with:
+- smaller block size / different benchmark settings
+- a shorter generation target
+- a different prompt sample
+- upstream MLX / Metal fixes
+- newer Apple Silicon hardware
+
+But as of this run, it should be treated as **experimental / failing** on this exact machine.
+
+## Recommendation
+
+For this Mac, the working local proof path is still:
+- `Qwen/Qwen3.5-4B`
+- `z-lab/Qwen3.5-4B-DFlash`
+
+Use the 4B pair for reproducible local validation while the 9B Metal timeout is investigated separately.

benchmarks/tool-call-regression.json

@@ -1,135 +0,0 @@
-{
-  "timestamp": "2026-04-16T01:56:48.462512+00:00",
-  "model": "dry-run",
-  "endpoint": "none",
-  "kv_type": "none",
-  "total": 10,
-  "passed": 10,
-  "failed": 0,
-  "accuracy": 1.0,
-  "meets_threshold": true,
-  "threshold": 1.0,
-  "results": [
-    {
-      "id": "read_file_basic",
-      "name": "Read File \u2014 basic path",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "read_file",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "read_file_offset",
-      "name": "Read File \u2014 with offset",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "read_file",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "web_search_basic",
-      "name": "Web Search \u2014 basic query",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "web_search",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "terminal_basic",
-      "name": "Terminal \u2014 simple command",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "terminal",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "terminal_complex",
-      "name": "Terminal \u2014 complex command",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "terminal",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "code_exec_basic",
-      "name": "Code Execution \u2014 python",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "execute_code",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "code_exec_complex",
-      "name": "Code Execution \u2014 multi-line",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "execute_code",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "delegate_basic",
-      "name": "Delegate Task \u2014 simple",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "delegate_task",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "delegate_context",
-      "name": "Delegate Task \u2014 with context",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "delegate_task",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    },
-    {
-      "id": "parallel_two",
-      "name": "Parallel Tools \u2014 two in one response",
-      "passed": true,
-      "tool_called": null,
-      "expected_tool": "read_file",
-      "schema_valid": true,
-      "args_valid": true,
-      "latency_ms": 0.0,
-      "raw_response": "",
-      "error": null
-    }
-  ],
-  "error": null
-}

benchmarks/tool-call-regression.md

@@ -1,32 +0,0 @@
-# Tool Call Regression Results
-
-**Generated:** 2026-04-16T01:56:48.462512+00:00
-**Model:** dry-run
-**Endpoint:** none
-**KV Type:** none
-
-## Summary
-
-| Metric | Value |
-|--------|-------|
-| Total tests | 10 |
-| Passed | 10 |
-| Failed | 0 |
-| Accuracy | 100.0% |
-| Threshold | 100% |
-| Verdict | PASS |
-
-## Test Matrix
-
-| Test ID | Tool Expected | Tool Called | Schema | Args | Latency | Status |
-|---------|--------------|-------------|--------|------|---------|--------|
-| read_file_basic | read_file | none | OK | OK | 0ms | PASS |
-| read_file_offset | read_file | none | OK | OK | 0ms | PASS |
-| web_search_basic | web_search | none | OK | OK | 0ms | PASS |
-| terminal_basic | terminal | none | OK | OK | 0ms | PASS |
-| terminal_complex | terminal | none | OK | OK | 0ms | PASS |
-| code_exec_basic | execute_code | none | OK | OK | 0ms | PASS |
-| code_exec_complex | execute_code | none | OK | OK | 0ms | PASS |
-| delegate_basic | delegate_task | none | OK | OK | 0ms | PASS |
-| delegate_context | delegate_task | none | OK | OK | 0ms | PASS |
-| parallel_two | read_file | none | OK | OK | 0ms | PASS |

docs/DFLASH_APPLE_SILICON.md

@@ -0,0 +1,125 @@
+# DFlash on Apple Silicon
+
+This repo now carries a **Gitea-first benchmark harness** for evaluating whether upstream **DFlash on MLX** is worth adding to the local Apple Silicon inference stack.
+
+## Why
+
+The headline `Kimi K2.6 + DFlash` benchmark was measured on `8x MI300X` with huge RAM and ROCm patches. That exact recipe is not a fit for a `36 GB` Apple Silicon Mac.
+
+What *is* relevant locally is the upstream `z-lab/dflash` MLX path, which can benchmark smaller matched target/draft pairs that fit on Apple Silicon.
+
+## Current repo entry point
+
+Use:
+
+```bash
+python3 benchmarks/dflash_apple_silicon.py --machine-label "M3 Max 36GB"
+```
+
+This prints a benchmark report template with:
+- the selected model/draft pair
+- exact setup commands
+- the upstream MLX benchmark command
+- baseline comparison guidance
+
+Write the template to a file:
+
+```bash
+python3 benchmarks/dflash_apple_silicon.py \
+  --machine-label "M3 Max 36GB" \
+  --output benchmarks/reports/dflash_m3max_36gb.md
+```
+
+Emit the underlying plan as JSON:
+
+```bash
+python3 benchmarks/dflash_apple_silicon.py --format json
+```
+
+## Selection logic
+
+Today the planner uses two upstream-supported MLX pairs:
+
+- `qwen35-9b`
+  - base: `Qwen/Qwen3.5-9B`
+  - draft: `z-lab/Qwen3.5-9B-DFlash`
+  - chosen for ~28 GB+ machines
+- `qwen35-4b`
+  - base: `Qwen/Qwen3.5-4B`
+  - draft: `z-lab/Qwen3.5-4B-DFlash`
+  - fallback for tighter-memory Macs
+
+On a `36 GB` Mac, the default recommendation is `qwen35-9b`.
+
+## Pilot result already landed
+
+A first live Apple Silicon run has already been captured in:
+
+- `benchmarks/reports/dflash_m3max_36gb_qwen35_4b_pilot.md`
+
+Pilot command:
+
+```bash
+python -m dflash.benchmark --backend mlx \
+    --model Qwen/Qwen3.5-4B \
+    --draft-model z-lab/Qwen3.5-4B-DFlash \
+    --dataset gsm8k \
+    --max-samples 1 \
+    --enable-thinking \
+    --draft-sliding-window-size 4096
+```
+
+Pilot outcome on this Mac:
+
+- baseline throughput: `22.35 tok/s`
+- DFlash throughput: `46.78 tok/s`
+- decoding speedup: `2.09x`
+
+Treat that as a **directional proof**, not a final decision benchmark. The next step is the fuller comparison slice against plain MLX or llama.cpp speculative decoding.
+
+## Known 9B failure on this machine
+
+A follow-up live run with:
+
+- `Qwen/Qwen3.5-9B`
+- `z-lab/Qwen3.5-9B-DFlash`
+
+failed on this same M3 Max 36GB Mac with:
+
+```text
+[METAL] Command buffer execution failed:
+Caused GPU Timeout Error (00000002:kIOGPUCommandBufferCallbackErrorTimeout)
+```
+
+That failure is recorded in:
+
+- `benchmarks/reports/dflash_m3max_36gb_qwen35_9b_timeout.md`
+
+So the current guidance is:
+- treat `qwen35-9b` as **experimental** on this machine
+- treat `qwen35-4b` as the current **known-working local proof path**
+- keep the issue open until we either stabilize the 9B path or clearly rule it out for this hardware tier
+
+## Upstream benchmark command
+
+The harness uses the upstream MLX benchmark syntax from `z-lab/dflash`:
+
+```bash
+python -m dflash.benchmark --backend mlx \
+    --model Qwen/Qwen3.5-9B \
+    --draft-model z-lab/Qwen3.5-9B-DFlash \
+    --dataset gsm8k \
+    --max-samples 128 \
+    --enable-thinking \
+    --draft-sliding-window-size 4096
+```
+
+## What remains
+
+This PR adds the **planner + report template** so the benchmark is reproducible from the repo.
+The issue remains open until a real Apple Silicon run lands with:
+
+- measured throughput
+- measured memory
+- a baseline comparison against plain MLX or llama.cpp speculative decoding
+- a recommendation on whether to operationalize DFlash locally

evolution/quant_selector.py

@@ -0,0 +1,548 @@
+"""Auto-select TurboQuant compression level based on available VRAM/RAM.
+
+Detects hardware resources at startup and picks the highest quality
+quantization level that fits within available memory. Supports Apple
+Silicon unified memory, NVIDIA GPUs (via nvidia-smi), and CPU-only fallback.
+
+Usage:
+    from evolution.quant_selector import select_quant_level
+
+    selection = select_quant_level(model_size_gb=14.0, context_length=32768)
+    print(selection.level)       # "turbo4"
+    print(selection.reasoning)   # "M4 Max 36GB unified: turbo4 fits 14.0GB model + ..."
+    print(selection.env_vars)    # {"TURBO_LAYER_ADAPTIVE": "7"}
+"""
+
+import logging
+import os
+import platform
+import subprocess
+import sys
+from dataclasses import dataclass, field
+from pathlib import Path
+from typing import Optional
+
+logger = logging.getLogger(__name__)
+
+
+# ── Quant Level Definitions ───────────────────────────────────────────────────
+
+@dataclass
+class QuantLevel:
+    """A TurboQuant compression level with its memory characteristics."""
+    name: str               # e.g. "turbo4"
+    bits_per_channel: float # e.g. 3.5 for turbo4
+    compression_ratio: float  # vs uncompressed KV cache
+    quality_label: str      # "best", "high", "balanced", "fast"
+    layer_adaptive: int     # TURBO_LAYER_ADAPTIVE value (0-7)
+    kv_type: str            # -ctk/-ctv flag value
+    min_memory_headroom_gb: float  # Minimum free memory to recommend this level
+    description: str = ""
+
+
+# Ordered from highest quality to most aggressive compression
+QUANT_LEVELS = [
+    QuantLevel(
+        name="turbo4",
+        bits_per_channel=3.5,
+        compression_ratio=4.2,
+        quality_label="best",
+        layer_adaptive=7,
+        kv_type="turbo4",
+        min_memory_headroom_gb=4.0,
+        description="PolarQuant + QJL 4-bit. Best quality, ~4.2x KV compression."
+    ),
+    QuantLevel(
+        name="turbo3",
+        bits_per_channel=2.5,
+        compression_ratio=6.0,
+        quality_label="high",
+        layer_adaptive=5,
+        kv_type="turbo3",
+        min_memory_headroom_gb=3.0,
+        description="3-bit TurboQuant. High quality, ~6x KV compression."
+    ),
+    QuantLevel(
+        name="turbo2",
+        bits_per_channel=1.5,
+        compression_ratio=10.0,
+        quality_label="balanced",
+        layer_adaptive=3,
+        kv_type="turbo2",
+        min_memory_headroom_gb=2.0,
+        description="2-bit TurboQuant. Balanced, ~10x KV compression."
+    ),
+    QuantLevel(
+        name="q4_0",
+        bits_per_channel=4.0,
+        compression_ratio=3.5,
+        quality_label="fast",
+        layer_adaptive=0,
+        kv_type="q4_0",
+        min_memory_headroom_gb=1.5,
+        description="Standard 4-bit quant. Fast fallback, no TurboQuant."
+    ),
+]
+
+
+# ── Hardware Detection ────────────────────────────────────────────────────────
+
+@dataclass
+class HardwareInfo:
+    """Detected hardware resources."""
+    total_memory_gb: float
+    available_memory_gb: float
+    gpu_memory_gb: Optional[float] = None
+    gpu_name: Optional[str] = None
+    is_apple_silicon: bool = False
+    chip_name: Optional[str] = None
+    cpu_cores: int = 0
+    detection_method: str = ""
+
+
+def detect_hardware() -> HardwareInfo:
+    """Detect available memory and GPU resources."""
+    system = platform.system()
+
+    if system == "Darwin":
+        return _detect_apple_silicon()
+    elif system == "Linux":
+        return _detect_linux()
+    else:
+        return _detect_generic(system)
+
+
+def _detect_apple_silicon() -> HardwareInfo:
+    """Detect Apple Silicon unified memory."""
+    info = HardwareInfo(
+        total_memory_gb=0,
+        available_memory_gb=0,
+        is_apple_silicon=True,
+        detection_method="sysctl",
+    )
+
+    try:
+        # Get total memory
+        result = subprocess.run(
+            ["sysctl", "-n", "hw.memsize"],
+            capture_output=True, text=True, timeout=5
+        )
+        if result.returncode == 0:
+            info.total_memory_gb = int(result.stdout.strip()) / (1024**3)
+
+        # Get chip name
+        result = subprocess.run(
+            ["sysctl", "-n", "machdep.cpu.brand_string"],
+            capture_output=True, text=True, timeout=5
+        )
+        if result.returncode == 0:
+            info.chip_name = result.stdout.strip()
+
+        # Try to get GPU name (Apple Silicon)
+        result = subprocess.run(
+            ["system_profiler", "SPDisplaysDataType"],
+            capture_output=True, text=True, timeout=10
+        )
+        if result.returncode == 0:
+            for line in result.stdout.split("\n"):
+                if "Chipset" in line or "GPU" in line:
+                    info.gpu_name = line.split(":")[-1].strip()
+                    break
+
+        # Estimate available memory (vm_stat)
+        result = subprocess.run(
+            ["vm_stat"],
+            capture_output=True, text=True, timeout=5
+        )
+        if result.returncode == 0:
+            page_size = 4096  # macOS default
+            free_pages = 0
+            for line in result.stdout.split("\n"):
+                if "Pages free:" in line:
+                    try:
+                        free_pages = int(line.split(":")[-1].strip().rstrip("."))
+                    except ValueError:
+                        pass
+            # Available ≈ free + some speculative (conservative: just free)
+            info.available_memory_gb = (free_pages * page_size) / (1024**3)
+
+        # Fallback if vm_stat parsing failed
+        if info.available_memory_gb < 1:
+            # Conservative: 70% of total
+            info.available_memory_gb = info.total_memory_gb * 0.70
+
+        # Apple Silicon shares memory — GPU memory = total memory
+        info.gpu_memory_gb = info.total_memory_gb
+
+        # Detect CPU cores
+        result = subprocess.run(
+            ["sysctl", "-n", "hw.ncpu"],
+            capture_output=True, text=True, timeout=5
+        )
+        if result.returncode == 0:
+            info.cpu_cores = int(result.stdout.strip())
+
+    except Exception as e:
+        logger.warning(f"Apple Silicon detection failed: {e}")
+        # Fallback
+        info.total_memory_gb = 16.0
+        info.available_memory_gb = 12.0
+        info.detection_method = "fallback"
+
+    return info
+
+
+def _detect_linux() -> HardwareInfo:
+    """Detect Linux system with optional NVIDIA GPU."""
+    info = HardwareInfo(
+        total_memory_gb=0,
+        available_memory_gb=0,
+        detection_method="proc",
+    )
+
+    try:
+        # Read /proc/meminfo
+        with open("/proc/meminfo", "r") as f:
+            meminfo = f.read()
+
+        for line in meminfo.split("\n"):
+            if line.startswith("MemTotal:"):
+                kb = int(line.split()[1])
+                info.total_memory_gb = kb / (1024 * 1024)
+            elif line.startswith("MemAvailable:"):
+                kb = int(line.split()[1])
+                info.available_memory_gb = kb / (1024 * 1024)
+
+        # CPU cores
+        info.cpu_cores = os.cpu_count() or 1
+
+        # Check for NVIDIA GPU
+        try:
+            result = subprocess.run(
+                ["nvidia-smi", "--query-gpu=name,memory.total,memory.free",
+                 "--format=csv,noheader,nounits"],
+                capture_output=True, text=True, timeout=10
+            )
+            if result.returncode == 0 and result.stdout.strip():
+                lines = result.stdout.strip().split("\n")
+                if lines:
+                    parts = lines[0].split(", ")
+                    if len(parts) >= 3:
+                        info.gpu_name = parts[0].strip()
+                        info.gpu_memory_gb = float(parts[1]) / 1024  # MB to GB
+                        gpu_free = float(parts[2]) / 1024
+                        # Use GPU free for VRAM-based selection
+                        info.available_memory_gb = max(info.available_memory_gb, gpu_free)
+                        info.detection_method = "nvidia-smi"
+        except (FileNotFoundError, subprocess.TimeoutExpired):
+            pass  # No NVIDIA GPU
+
+    except Exception as e:
+        logger.warning(f"Linux detection failed: {e}")
+        info.total_memory_gb = 16.0
+        info.available_memory_gb = 12.0
+        info.detection_method = "fallback"
+
+    return info
+
+
+def _detect_generic(system: str) -> HardwareInfo:
+    """Fallback detection for unknown systems."""
+    import psutil
+    mem = psutil.virtual_memory()
+    return HardwareInfo(
+        total_memory_gb=mem.total / (1024**3),
+        available_memory_gb=mem.available / (1024**3),
+        cpu_cores=os.cpu_count() or 1,
+        detection_method="psutil",
+    )
+
+
+# ── KV Cache Memory Estimation ───────────────────────────────────────────────
+
+def estimate_kv_cache_gb(
+    context_length: int,
+    num_layers: int = 48,
+    num_kv_heads: int = 8,
+    head_dim: int = 128,
+    bits_per_channel: float = 3.5,
+) -> float:
+    """Estimate KV cache memory for given parameters.
+
+    Formula: 2 (K+V) × layers × kv_heads × head_dim × context_length × bits/8
+    """
+    bytes_per_element = bits_per_channel / 8.0
+    total_bytes = 2 * num_layers * num_kv_heads * head_dim * context_length * bytes_per_element
+    return total_bytes / (1024**3)
+
+
+def estimate_model_memory_gb(model_size_gb: float, quant_type: str = "q4_k_m") -> float:
+    """Estimate model weights memory. Returns loaded size in GB.
+
+    This is a rough estimate — actual depends on exact quant format.
+    """
+    # Common quant ratios (vs fp16)
+    quant_multipliers = {
+        "f16": 1.0,
+        "q8_0": 0.5,
+        "q6_k": 0.42,
+        "q5_k_m": 0.37,
+        "q4_k_m": 0.32,
+        "q3_k_m": 0.27,
+        "q2_k": 0.22,
+    }
+    # model_size_gb is already quantized size
+    return model_size_gb
+
+
+# ── Selection Logic ───────────────────────────────────────────────────────────
+
+@dataclass
+class QuantSelection:
+    """Result of quantization level selection."""
+    level: QuantLevel
+    hardware: HardwareInfo
+    reasoning: str
+    total_required_gb: float
+    available_gb: float
+    headroom_gb: float
+    env_vars: dict = field(default_factory=dict)
+    server_flags: dict = field(default_factory=dict)
+    warnings: list = field(default_factory=list)
+
+
+def select_quant_level(
+    model_size_gb: float = 14.0,
+    context_length: int = 32768,
+    num_layers: int = 48,
+    num_kv_heads: int = 8,
+    head_dim: int = 128,
+    preferred_level: Optional[str] = None,
+    force_cpu: bool = False,
+) -> QuantSelection:
+    """Select the best quantization level for available hardware.
+
+    Args:
+        model_size_gb: Size of the model weights in GB
+        context_length: Target context length
+        num_layers: Number of transformer layers
+        num_kv_heads: Number of KV attention heads
+        head_dim: Dimension per attention head
+        preferred_level: Force a specific level (still checks if it fits)
+        force_cpu: If True, ignore GPU memory
+
+    Returns:
+        QuantSelection with the chosen level and reasoning
+    """
+    hw = detect_hardware()
+
+    if force_cpu:
+        hw.gpu_memory_gb = None
+        hw.gpu_name = None
+
+    # Use the most restrictive memory constraint
+    # For Apple Silicon: unified memory, use total
+    # For NVIDIA: use GPU VRAM
+    # For CPU-only: use system RAM
+    if hw.gpu_memory_gb and hw.gpu_name:
+        memory_pool_gb = hw.gpu_memory_gb
+        memory_label = f"{hw.gpu_name} {hw.gpu_memory_gb:.0f}GB VRAM"
+    elif hw.is_apple_silicon:
+        memory_pool_gb = hw.total_memory_gb
+        memory_label = f"{hw.chip_name or 'Apple Silicon'} {hw.total_memory_gb:.0f}GB unified"
+    else:
+        memory_pool_gb = hw.total_memory_gb
+        memory_label = f"{hw.cpu_cores}c CPU {hw.total_memory_gb:.0f}GB RAM"
+
+    model_mem = estimate_model_memory_gb(model_size_gb)
+
+    # Try levels from best to most compressed
+    chosen = None
+    for level in QUANT_LEVELS:
+        if preferred_level and level.name != preferred_level:
+            continue
+
+        kv_mem = estimate_kv_cache_gb(
+            context_length, num_layers, num_kv_heads, head_dim,
+            level.bits_per_channel
+        )
+        total_required = model_mem + kv_mem
+        headroom = memory_pool_gb - total_required
+
+        if headroom >= level.min_memory_headroom_gb:
+            chosen = level
+            break
+
+        if preferred_level and level.name == preferred_level:
+            # User forced this level but it doesn't fit
+            chosen = level
+            break
+
+    if chosen is None:
+        # Nothing fits — pick the most aggressive compression, not the q4_0 fallback.
+        chosen = max(QUANT_LEVELS, key=lambda level: level.compression_ratio)
+        logger.warning(f"No quant level fits in {memory_pool_gb:.1f}GB. Using {chosen.name}.")
+
+    # Calculate final numbers
+    kv_mem = estimate_kv_cache_gb(
+        context_length, num_layers, num_kv_heads, head_dim,
+        chosen.bits_per_channel
+    )
+    total_required = model_mem + kv_mem
+    headroom = memory_pool_gb - total_required
+
+    # Build reasoning
+    reasoning_parts = [
+        f"{memory_label}:",
+        f"{chosen.name} ({chosen.quality_label}, {chosen.bits_per_channel:.1f}b/ch,",
+        f"{chosen.compression_ratio:.1f}x compression)",
+        f"fits {model_mem:.1f}GB model + {kv_mem:.1f}GB KV cache",
+        f"@ {context_length}K context = {total_required:.1f}GB / {memory_pool_gb:.0f}GB",
+        f"({headroom:.1f}GB headroom)"
+    ]
+    reasoning = " ".join(reasoning_parts)
+
+    # Build environment variables for llama.cpp
+    env_vars = {
+        "TURBO_LAYER_ADAPTIVE": str(chosen.layer_adaptive),
+    }
+
+    # Build server flags
+    server_flags = {
+        "-ctk": chosen.kv_type,
+        "-ctv": chosen.kv_type,
+        "-c": str(context_length),
+    }
+
+    # Warnings
+    warnings = []
+    if headroom < 2.0:
+        warnings.append(
+            f"Low headroom ({headroom:.1f}GB). Consider reducing context length or model size."
+        )
+    if headroom < 0:
+        warnings.append(
+            f"OVERCOMMITTED: needs {total_required:.1f}GB but only {memory_pool_gb:.0f}GB available. "
+            f"Inference may fail or swap heavily."
+        )
+
+    selection = QuantSelection(
+        level=chosen,
+        hardware=hw,
+        reasoning=reasoning,
+        total_required_gb=total_required,
+        available_gb=memory_pool_gb,
+        headroom_gb=headroom,
+        env_vars=env_vars,
+        server_flags=server_flags,
+        warnings=warnings,
+    )
+
+    logger.info(f"Quant selection: {reasoning}")
+    for w in warnings:
+        logger.warning(w)
+
+    return selection
+
+
+# ── CLI ───────────────────────────────────────────────────────────────────────
+
+def main():
+    """CLI entry point for quant level selection."""
+    import argparse
+    import json
+
+    parser = argparse.ArgumentParser(
+        description="Auto-select TurboQuant compression level based on available hardware"
+    )
+    parser.add_argument("--model-size", type=float, default=14.0,
+                        help="Model size in GB (default: 14.0)")
+    parser.add_argument("--context", type=int, default=32768,
+                        help="Target context length (default: 32768)")
+    parser.add_argument("--layers", type=int, default=48,
+                        help="Number of transformer layers (default: 48)")
+    parser.add_argument("--kv-heads", type=int, default=8,
+                        help="Number of KV attention heads (default: 8)")
+    parser.add_argument("--head-dim", type=int, default=128,
+                        help="Dimension per attention head (default: 128)")
+    parser.add_argument("--prefer", type=str, default=None,
+                        choices=[l.name for l in QUANT_LEVELS],
+                        help="Prefer a specific quant level")
+    parser.add_argument("--force-cpu", action="store_true",
+                        help="Ignore GPU, use CPU memory only")
+    parser.add_argument("--json", action="store_true",
+                        help="JSON output for automation")
+    parser.add_argument("--detect-only", action="store_true",
+                        help="Only detect hardware, don't select")
+    args = parser.parse_args()
+
+    logging.basicConfig(level=logging.INFO, format="%(message)s")
+
+    if args.detect_only:
+        hw = detect_hardware()
+        if args.json:
+            print(json.dumps(hw.__dict__, default=str, indent=2))
+        else:
+            print(f"Total memory:    {hw.total_memory_gb:.1f} GB")
+            print(f"Available:       {hw.available_memory_gb:.1f} GB")
+            if hw.gpu_memory_gb:
+                print(f"GPU memory:      {hw.gpu_memory_gb:.1f} GB")
+            if hw.gpu_name:
+                print(f"GPU:             {hw.gpu_name}")
+            if hw.is_apple_silicon:
+                print(f"Chip:            {hw.chip_name or 'Apple Silicon'}")
+            print(f"CPU cores:       {hw.cpu_cores}")
+            print(f"Detection:       {hw.detection_method}")
+        return
+
+    selection = select_quant_level(
+        model_size_gb=args.model_size,
+        context_length=args.context,
+        num_layers=args.layers,
+        num_kv_heads=args.kv_heads,
+        head_dim=args.head_dim,
+        preferred_level=args.prefer,
+        force_cpu=args.force_cpu,
+    )
+
+    if args.json:
+        result = {
+            "level": selection.level.name,
+            "bits_per_channel": selection.level.bits_per_channel,
+            "compression_ratio": selection.level.compression_ratio,
+            "quality": selection.level.quality_label,
+            "reasoning": selection.reasoning,
+            "total_required_gb": round(selection.total_required_gb, 2),
+            "available_gb": round(selection.available_gb, 1),
+            "headroom_gb": round(selection.headroom_gb, 2),
+            "env_vars": selection.env_vars,
+            "server_flags": selection.server_flags,
+            "warnings": selection.warnings,
+            "hardware": {
+                "total_memory_gb": round(selection.hardware.total_memory_gb, 1),
+                "gpu_name": selection.hardware.gpu_name,
+                "is_apple_silicon": selection.hardware.is_apple_silicon,
+                "chip_name": selection.hardware.chip_name,
+                "cpu_cores": selection.hardware.cpu_cores,
+            },
+        }
+        print(json.dumps(result, indent=2))
+    else:
+        print(f"Selected: {selection.level.name} ({selection.level.quality_label})")
+        print(f"  {selection.reasoning}")
+        print()
+        print(f"Environment variables:")
+        for k, v in selection.env_vars.items():
+            print(f"  export {k}={v}")
+        print()
+        print(f"Server flags:")
+        for k, v in selection.server_flags.items():
+            print(f"  {k} {v}")
+        if selection.warnings:
+            print()
+            for w in selection.warnings:
+                print(f"  WARNING: {w}")
+
+
+if __name__ == "__main__":
+    main()

tests/conftest.py

@@ -0,0 +1,3 @@
+"""Pytest configuration for turboquant."""
+import sys, os
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))

tests/test_dflash_apple_silicon.py

@@ -0,0 +1,58 @@
+#!/usr/bin/env python3
+"""Tests for Apple Silicon DFlash benchmark planning helpers (issue #152)."""
+
+import os
+import sys
+from unittest.mock import patch
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(__file__)))
+
+from benchmarks.dflash_apple_silicon import (  # noqa: E402
+    build_mlx_benchmark_command,
+    detect_total_memory_gb,
+    render_report_template,
+    select_pair,
+)
+
+
+class TestPairSelection:
+    def test_prefers_qwen35_9b_on_36gb_mac(self):
+        pair = select_pair(total_memory_gb=36)
+        assert pair.slug == "qwen35-9b"
+        assert pair.base_model == "Qwen/Qwen3.5-9B"
+        assert pair.draft_model == "z-lab/Qwen3.5-9B-DFlash"
+
+    def test_falls_back_to_4b_when_memory_is_tight(self):
+        pair = select_pair(total_memory_gb=20)
+        assert pair.slug == "qwen35-4b"
+        assert pair.base_model == "Qwen/Qwen3.5-4B"
+
+
+class TestCommandGeneration:
+    def test_builds_upstream_mlx_benchmark_command(self):
+        pair = select_pair(total_memory_gb=36)
+        command = build_mlx_benchmark_command(pair, dataset="gsm8k", max_samples=64)
+        assert "python -m dflash.benchmark --backend mlx" in command
+        assert "--model Qwen/Qwen3.5-9B" in command
+        assert "--draft-model z-lab/Qwen3.5-9B-DFlash" in command
+        assert "--dataset gsm8k" in command
+        assert "--max-samples 64" in command
+        assert "--draft-sliding-window-size 4096" in command
+
+
+class TestReportTemplate:
+    def test_report_template_mentions_baseline_and_verdict(self):
+        pair = select_pair(total_memory_gb=36)
+        report = render_report_template(machine_label="M3 Max 36GB", pair=pair)
+        assert "DFlash Apple Silicon Benchmark Report" in report
+        assert "M3 Max 36GB" in report
+        assert "Qwen/Qwen3.5-9B" in report
+        assert "plain MLX or llama.cpp speculative decoding" in report
+        assert "Worth operationalizing locally?" in report
+
+
+class TestMemoryDetection:
+    @patch("benchmarks.dflash_apple_silicon.platform.system", return_value="Darwin")
+    @patch("benchmarks.dflash_apple_silicon.subprocess.check_output", return_value=b"38654705664\n")
+    def test_detect_total_memory_gb_on_macos(self, _mock_sysctl, _mock_system):
+        assert detect_total_memory_gb() == 36.0

tests/test_quant_selector.py

@@ -0,0 +1,177 @@
+#!/usr/bin/env python3
+"""Tests for quant_selector.py"""
+
+import sys
+import os
+import pytest
+from unittest.mock import patch, MagicMock
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(__file__)))
+from evolution.quant_selector import (
+    QuantLevel,
+    HardwareInfo,
+    QUANT_LEVELS,
+    detect_hardware,
+    estimate_kv_cache_gb,
+    estimate_model_memory_gb,
+    select_quant_level,
+)
+
+
+class TestQuantLevels:
+    def test_levels_keep_turboquant_quality_order_with_q4_fallback_last(self):
+        """TurboQuant levels should lead, with q4_0 reserved as the non-Turbo fallback."""
+        names = [level.name for level in QUANT_LEVELS]
+        assert names[:3] == ["turbo4", "turbo3", "turbo2"]
+        assert names[-1] == "q4_0"
+
+    def test_all_levels_have_required_fields(self):
+        for level in QUANT_LEVELS:
+            assert level.name
+            assert level.bits_per_channel > 0
+            assert level.compression_ratio > 1
+            assert level.quality_label
+            assert level.layer_adaptive >= 0
+            assert level.kv_type
+
+
+class TestKVEstimate:
+    def test_basic_estimate(self):
+        # 48 layers, 8 heads, 128 dim, 32K context, 3.5 bits
+        kv_gb = estimate_kv_cache_gb(32768, 48, 8, 128, 3.5)
+        assert kv_gb > 0
+        assert kv_gb < 10  # Should be reasonable
+
+    def test_longer_context_larger(self):
+        kv_32k = estimate_kv_cache_gb(32768, 48, 8, 128, 3.5)
+        kv_128k = estimate_kv_cache_gb(131072, 48, 8, 128, 3.5)
+        assert kv_128k > kv_32k
+
+    def test_higher_bits_larger(self):
+        kv_4b = estimate_kv_cache_gb(32768, 48, 8, 128, 4.0)
+        kv_2b = estimate_kv_cache_gb(32768, 48, 8, 128, 2.0)
+        assert kv_4b > kv_2b
+
+
+class TestHardwareDetection:
+    def test_detect_returns_info(self):
+        hw = detect_hardware()
+        assert hw.total_memory_gb > 0
+        assert hw.available_memory_gb > 0
+        assert hw.detection_method
+
+    @patch("evolution.quant_selector.platform.system", return_value="Linux")
+    @patch("builtins.open", create=True)
+    def test_linux_detection(self, mock_open, mock_system):
+        mock_open.return_value.__enter__().read.return_value = (
+            "MemTotal:       32000000 kB\n"
+            "MemAvailable:   24000000 kB\n"
+        )
+        hw = _detect_linux_fallback()
+        assert hw.total_memory_gb > 20
+
+
+def _detect_linux_fallback():
+    """Helper to test Linux detection with mocked /proc/meminfo."""
+    from evolution.quant_selector import _detect_linux
+    return _detect_linux()
+
+
+class TestSelection:
+    def test_selects_turbo4_for_large_memory(self):
+        """With plenty of memory, should pick turbo4 (best quality)."""
+        with patch("evolution.quant_selector.detect_hardware") as mock_hw:
+            mock_hw.return_value = HardwareInfo(
+                total_memory_gb=64,
+                available_memory_gb=48,
+                gpu_memory_gb=64,
+                gpu_name="Test GPU",
+                cpu_cores=16,
+                detection_method="mock",
+            )
+            sel = select_quant_level(model_size_gb=14.0, context_length=32768)
+            assert sel.level.name == "turbo4"
+            assert sel.headroom_gb > 0
+
+    def test_selects_smaller_for_tight_memory(self):
+        """With tight memory, should pick a smaller quant."""
+        with patch("evolution.quant_selector.detect_hardware") as mock_hw:
+            mock_hw.return_value = HardwareInfo(
+                total_memory_gb=16,
+                available_memory_gb=12,
+                gpu_memory_gb=16,
+                gpu_name="Test GPU",
+                cpu_cores=8,
+                detection_method="mock",
+            )
+            sel = select_quant_level(model_size_gb=14.0, context_length=131072)
+            # Should pick a smaller quant for 128K context on 16GB
+            assert sel.level.bits_per_channel <= 4.0
+
+    def test_preferred_level(self):
+        """User can force a specific level."""
+        with patch("evolution.quant_selector.detect_hardware") as mock_hw:
+            mock_hw.return_value = HardwareInfo(
+                total_memory_gb=64,
+                available_memory_gb=48,
+                cpu_cores=16,
+                detection_method="mock",
+            )
+            sel = select_quant_level(
+                model_size_gb=14.0, context_length=32768,
+                preferred_level="turbo2"
+            )
+            assert sel.level.name == "turbo2"
+
+    def test_env_vars_populated(self):
+        with patch("evolution.quant_selector.detect_hardware") as mock_hw:
+            mock_hw.return_value = HardwareInfo(
+                total_memory_gb=64,
+                available_memory_gb=48,
+                cpu_cores=16,
+                detection_method="mock",
+            )
+            sel = select_quant_level(model_size_gb=14.0, context_length=32768)
+            assert "TURBO_LAYER_ADAPTIVE" in sel.env_vars
+            assert "-ctk" in sel.server_flags
+            assert "-ctv" in sel.server_flags
+
+    def test_warnings_on_low_headroom(self):
+        with patch("evolution.quant_selector.detect_hardware") as mock_hw:
+            mock_hw.return_value = HardwareInfo(
+                total_memory_gb=18,
+                available_memory_gb=14,
+                gpu_memory_gb=18,
+                gpu_name="Test GPU",
+                cpu_cores=8,
+                detection_method="mock",
+            )
+            sel = select_quant_level(model_size_gb=16.0, context_length=65536)
+            assert len(sel.warnings) > 0
+
+    def test_falls_back_to_turbo2_when_nothing_fits(self):
+        with patch("evolution.quant_selector.detect_hardware") as mock_hw:
+            mock_hw.return_value = HardwareInfo(
+                total_memory_gb=8,
+                available_memory_gb=6,
+                gpu_memory_gb=8,
+                gpu_name="Tiny GPU",
+                cpu_cores=4,
+                detection_method="mock",
+            )
+            sel = select_quant_level(model_size_gb=16.0, context_length=131072)
+            assert sel.level.name == "turbo2"
+
+    def test_reasoning_contains_key_info(self):
+        with patch("evolution.quant_selector.detect_hardware") as mock_hw:
+            mock_hw.return_value = HardwareInfo(
+                total_memory_gb=32,
+                available_memory_gb=24,
+                is_apple_silicon=True,
+                chip_name="M4 Max",
+                cpu_cores=16,
+                detection_method="mock",
+            )
+            sel = select_quant_level(model_size_gb=14.0, context_length=32768)
+            assert "turbo4" in sel.reasoning
+            assert "M4 Max" in sel.reasoning or "32GB" in sel.reasoning

tests/test_tool_call_integration.py

@@ -0,0 +1,338 @@
+"""
+Integration test: turboquant compressed model passes hermes tool calls (issue #82).
+
+Validates that a TurboQuant-compressed model can:
+1. Parse hermes tool schemas correctly
+2. Format tool calls in OpenAI-compatible format
+3. Pass through the hermes agent conversation loop
+
+Tests are structured as contract tests -- they validate the schema/format
+compatibility without requiring a running model server. The live inference
+test is skipped by default (requires llama-server with TurboQuant model).
+
+Usage:
+    pytest tests/test_tool_call_integration.py -v
+    pytest tests/test_tool_call_integration.py -v -k live  # run live test if server available
+"""
+import json
+import os
+import pathlib
+import re
+import unittest
+
+import pytest
+
+ROOT = pathlib.Path(__file__).resolve().parents[1]
+PROFILE_PATH = ROOT / "profiles" / "hermes-profile-gemma4-turboquant.yaml"
+BENCHMARKS_DIR = ROOT / "benchmarks"
+
+
+class TestHermesProfileSchema(unittest.TestCase):
+    """Validate the hermes profile YAML has required fields for tool calling."""
+
+    @classmethod
+    def setUpClass(cls):
+        import yaml
+        cls.profile = yaml.safe_load(PROFILE_PATH.read_text())
+
+    def test_profile_has_providers(self):
+        assert "providers" in self.profile, "Profile must define providers"
+        assert "primary" in self.profile["providers"], "Must have primary provider"
+
+    def test_primary_provider_has_endpoint(self):
+        primary = self.profile["providers"]["primary"]
+        assert "endpoint" in primary, "Primary provider must have endpoint"
+        assert primary["endpoint"].startswith("http"), "Endpoint must be HTTP(S) URL"
+
+    def test_primary_provider_has_api_path(self):
+        primary = self.profile["providers"]["primary"]
+        assert "api_path" in primary, "Primary provider must have api_path"
+        assert "/chat/completions" in primary["api_path"], (
+            "api_path should be OpenAI-compatible /chat/completions"
+        )
+
+    def test_turboquant_settings_present(self):
+        primary = self.profile["providers"]["primary"]
+        assert "turboquant" in primary, "Must have turboquant config section"
+        tq = primary["turboquant"]
+        assert tq.get("enabled") is True, "TurboQuant must be enabled"
+        assert tq.get("kv_type") in ("turbo2", "turbo3", "turbo4"), (
+            "kv_type must be turbo2, turbo3, or turbo4"
+        )
+
+    def test_context_window_configured(self):
+        primary = self.profile["providers"]["primary"]
+        assert "context" in primary, "Must have context config"
+        ctx = primary["context"]
+        assert ctx.get("max_tokens", 0) >= 8192, (
+            "max_tokens should be >= 8192 for TurboQuant value proposition"
+        )
+
+
+class TestToolSchemaCompatibility(unittest.TestCase):
+    """Verify hermes tool schemas serialize to valid JSON for OpenAI tool_calls."""
+
+    SAMPLE_TOOL_SCHEMAS = [
+        {
+            "type": "function",
+            "function": {
+                "name": "read_file",
+                "description": "Read a text file with line numbers.",
+                "parameters": {
+                    "type": "object",
+                    "properties": {
+                        "path": {"type": "string", "description": "File path"},
+                        "offset": {"type": "integer", "default": 1},
+                        "limit": {"type": "integer", "default": 500},
+                    },
+                    "required": ["path"],
+                },
+            },
+        },
+        {
+            "type": "function",
+            "function": {
+                "name": "execute_code",
+                "description": "Run a Python script.",
+                "parameters": {
+                    "type": "object",
+                    "properties": {
+                        "code": {"type": "string", "description": "Python code"},
+                    },
+                    "required": ["code"],
+                },
+            },
+        },
+        {
+            "type": "function",
+            "function": {
+                "name": "web_search",
+                "description": "Search the web.",
+                "parameters": {
+                    "type": "object",
+                    "properties": {
+                        "query": {"type": "string"},
+                        "max_results": {"type": "integer", "default": 5},
+                    },
+                    "required": ["query"],
+                },
+            },
+        },
+    ]
+
+    def test_tool_schemas_serialize_to_json(self):
+        """Tool schemas must serialize without errors."""
+        serialized = json.dumps(self.SAMPLE_TOOL_SCHEMAS)
+        assert len(serialized) > 0
+        parsed = json.loads(serialized)
+        assert len(parsed) == len(self.SAMPLE_TOOL_SCHEMAS)
+
+    def test_tool_schemas_have_required_openai_fields(self):
+        """Each tool schema must have the fields OpenAI expects."""
+        for tool in self.SAMPLE_TOOL_SCHEMAS:
+            assert tool["type"] == "function", "Tool type must be 'function'"
+            fn = tool["function"]
+            assert "name" in fn, "Function must have name"
+            assert "description" in fn, "Function must have description"
+            assert "parameters" in fn, "Function must have parameters"
+            params = fn["parameters"]
+            assert params["type"] == "object", "Parameters type must be 'object'"
+            assert "properties" in params, "Parameters must have properties"
+
+    def test_tool_call_response_format(self):
+        """Verify tool_call response matches OpenAI format."""
+        tool_call = {
+            "id": "call_abc123",
+            "type": "function",
+            "function": {
+                "name": "read_file",
+                "arguments": json.dumps({"path": "/tmp/test.txt"}),
+            },
+        }
+        args = json.loads(tool_call["function"]["arguments"])
+        assert args["path"] == "/tmp/test.txt"
+        assert tool_call["function"]["name"] in [
+            t["function"]["name"] for t in self.SAMPLE_TOOL_SCHEMAS
+        ]
+
+    def test_tool_names_are_valid_identifiers(self):
+        """Tool names must be valid Python identifiers for hermes dispatch."""
+        for tool in self.SAMPLE_TOOL_SCHEMAS:
+            name = tool["function"]["name"]
+            assert re.match(r"^[a-zA-Z_][a-zA-Z0-9_]*$", name), (
+                f"Tool name \'{name}\' is not a valid identifier"
+            )
+
+
+class TestTurboquantServerConfig(unittest.TestCase):
+    """Validate server startup configuration matches hermes profile."""
+
+    def test_server_command_has_turboquant_flags(self):
+        """The server command in the profile must include -ctk/-ctv flags."""
+        profile_text = PROFILE_PATH.read_text()
+        assert "-ctk" in profile_text, "Profile server command must include -ctk flag"
+        assert "-ctv" in profile_text, "Profile server command must include -ctv flag"
+
+    def test_server_command_has_context_flag(self):
+        """Server command must set context size."""
+        profile_text = PROFILE_PATH.read_text()
+        assert re.search(r"-c\s+\d+", profile_text), (
+            "Server command must include -c <context_size> flag"
+        )
+
+    def test_layer_adaptive_env_var(self):
+        """Profile must set TURBO_LAYER_ADAPTIVE env var."""
+        profile_text = PROFILE_PATH.read_text()
+        assert "TURBO_LAYER_ADAPTIVE" in profile_text, (
+            "Profile must configure TURBO_LAYER_ADAPTIVE"
+        )
+
+
+class TestBenchmarkData(unittest.TestCase):
+    """Validate benchmark test prompts include tool-call test cases."""
+
+    @classmethod
+    def setUpClass(cls):
+        prompts_path = BENCHMARKS_DIR / "test_prompts.json"
+        cls.prompts = json.loads(prompts_path.read_text())
+
+    def test_has_tool_call_test_prompt(self):
+        """Benchmark prompts must include a tool-call format test."""
+        categories = [p.get("category") for p in self.prompts]
+        assert "tool_call_format" in categories, (
+            "Benchmark must include a tool_call_format test case"
+        )
+
+    def test_tool_call_prompt_expects_json(self):
+        """Tool call test prompt must expect JSON in the response."""
+        tool_prompt = next(
+            p for p in self.prompts if p.get("category") == "tool_call_format"
+        )
+        pattern = tool_prompt.get("expected_pattern", "")
+        assert "json" in pattern.lower() or "\\{" in pattern, (
+            "Tool call prompt must expect JSON-formatted response"
+        )
+
+
+@pytest.mark.skipif(
+    not os.environ.get("TURBOQUANT_SERVER_URL"),
+    reason="No TurboQuant server available (set TURBOQUANT_SERVER_URL to run)",
+)
+class TestLiveToolCallIntegration:
+    """Live integration test -- requires running llama-server with TurboQuant."""
+
+    def test_server_health(self):
+        """Server must respond to /v1/models endpoint."""
+        import requests
+        url = os.environ["TURBOQUANT_SERVER_URL"]
+        resp = requests.get(f"{url}/v1/models", timeout=10)
+        assert resp.status_code == 200
+        data = resp.json()
+        assert "data" in data
+        assert len(data["data"]) > 0
+
+    def test_tool_call_completion(self):
+        """Model must return a valid tool_call for a read_file prompt."""
+        import requests
+        url = os.environ["TURBOQUANT_SERVER_URL"]
+        tools = [
+            {
+                "type": "function",
+                "function": {
+                    "name": "read_file",
+                    "description": "Read a file",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {"path": {"type": "string"}},
+                        "required": ["path"],
+                    },
+                },
+            }
+        ]
+        resp = requests.post(
+            f"{url}/v1/chat/completions",
+            json={
+                "model": "gemma-4",
+                "messages": [
+                    {"role": "user", "content": "Read the file at /tmp/test.txt"}
+                ],
+                "tools": tools,
+                "tool_choice": "auto",
+            },
+            timeout=120,
+        )
+        assert resp.status_code == 200
+        data = resp.json()
+        choice = data["choices"][0]
+        msg = choice["message"]
+        if "tool_calls" in msg and msg["tool_calls"]:
+            tc = msg["tool_calls"][0]
+            assert tc["type"] == "function"
+            assert tc["function"]["name"] == "read_file"
+            args = json.loads(tc["function"]["arguments"])
+            assert "path" in args
+        else:
+            assert len(msg.get("content", "")) > 0
+
+    def test_tool_call_with_multiple_tools(self):
+        """Model must handle multiple available tools."""
+        import requests
+        url = os.environ["TURBOQUANT_SERVER_URL"]
+        tools = [
+            {
+                "type": "function",
+                "function": {
+                    "name": "read_file",
+                    "description": "Read a file",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {"path": {"type": "string"}},
+                        "required": ["path"],
+                    },
+                },
+            },
+            {
+                "type": "function",
+                "function": {
+                    "name": "web_search",
+                    "description": "Search the web",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {"query": {"type": "string"}},
+                        "required": ["query"],
+                    },
+                },
+            },
+            {
+                "type": "function",
+                "function": {
+                    "name": "execute_code",
+                    "description": "Run Python code",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {"code": {"type": "string"}},
+                        "required": ["code"],
+                    },
+                },
+            },
+        ]
+        resp = requests.post(
+            f"{url}/v1/chat/completions",
+            json={
+                "model": "gemma-4",
+                "messages": [
+                    {"role": "user", "content": "Search the web for 'bitcoin price'"}
+                ],
+                "tools": tools,
+                "tool_choice": "auto",
+            },
+            timeout=120,
+        )
+        assert resp.status_code == 200
+        data = resp.json()
+        assert "choices" in data
+        assert len(data["choices"]) > 0
+
+
+if __name__ == "__main__":
+    unittest.main()

tests/tool_call_regression.py

@@ -1,678 +0,0 @@
-#!/usr/bin/env python3
-"""
-TurboQuant Tool Call Regression Suite (Issue #96)
-
-Verifies that TurboQuant-compressed models still handle hermes tool calling
-correctly. Tests schema parsing, execution, and parallel tool calls.
-
-Usage:
-    python3 tests/tool_call_regression.py \
-        --endpoint http://localhost:8081/v1 \
-        --model gemma-4 \
-        --kv-type turbo4 \
-        --runs 3
-
-    # Dry run (no server needed — validates schemas only):
-    python3 tests/tool_call_regression.py --dry-run
-
-Acceptance: tool call accuracy must be >= 95% across all test cases.
-"""
-
-import argparse
-import json
-import os
-import re
-import sys
-import time
-from dataclasses import dataclass, field, asdict
-from datetime import datetime, timezone
-from typing import Optional
-
-
-# ── Tool schemas (hermes-compatible) ──────────────────────────────
-
-TOOL_SCHEMAS = [
-    {
-        "type": "function",
-        "function": {
-            "name": "read_file",
-            "description": "Read a text file with line numbers and pagination.",
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "path": {
-                        "type": "string",
-                        "description": "File path to read (absolute or relative)"
-                    },
-                    "offset": {
-                        "type": "integer",
-                        "description": "Line number to start reading from (1-indexed)",
-                        "default": 1
-                    },
-                    "limit": {
-                        "type": "integer",
-                        "description": "Maximum number of lines to return",
-                        "default": 500
-                    }
-                },
-                "required": ["path"]
-            }
-        }
-    },
-    {
-        "type": "function",
-        "function": {
-            "name": "web_search",
-            "description": "Search the web for information using a query string.",
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "query": {
-                        "type": "string",
-                        "description": "Search query"
-                    },
-                    "num_results": {
-                        "type": "integer",
-                        "description": "Number of results to return",
-                        "default": 5
-                    }
-                },
-                "required": ["query"]
-            }
-        }
-    },
-    {
-        "type": "function",
-        "function": {
-            "name": "terminal",
-            "description": "Execute a shell command on the system.",
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "command": {
-                        "type": "string",
-                        "description": "Shell command to execute"
-                    },
-                    "timeout": {
-                        "type": "integer",
-                        "description": "Timeout in seconds",
-                        "default": 30
-                    }
-                },
-                "required": ["command"]
-            }
-        }
-    },
-    {
-        "type": "function",
-        "function": {
-            "name": "execute_code",
-            "description": "Run a Python script in a sandboxed environment.",
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "code": {
-                        "type": "string",
-                        "description": "Python code to execute"
-                    }
-                },
-                "required": ["code"]
-            }
-        }
-    },
-    {
-        "type": "function",
-        "function": {
-            "name": "delegate_task",
-            "description": "Spawn a subagent to work on a task in an isolated context.",
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "goal": {
-                        "type": "string",
-                        "description": "What the subagent should accomplish"
-                    },
-                    "context": {
-                        "type": "string",
-                        "description": "Background information the subagent needs"
-                    },
-                    "toolsets": {
-                        "type": "array",
-                        "items": {"type": "string"},
-                        "description": "Toolsets to enable for this subagent"
-                    }
-                },
-                "required": ["goal"]
-            }
-        }
-    },
-]
-
-
-# ── Test prompts ──────────────────────────────────────────────────
-
-@dataclass
-class ToolCallTest:
-    """A single test case for tool calling."""
-    id: str
-    name: str
-    prompt: str
-    expected_tool: str
-    expected_args: dict  # subset of expected args
-    description: str = ""
-
-
-TEST_CASES = [
-    ToolCallTest(
-        id="read_file_basic",
-        name="Read File — basic path",
-        prompt="Read the file at /tmp/test.txt and show me the first 10 lines.",
-        expected_tool="read_file",
-        expected_args={"path": "/tmp/test.txt"},
-        description="Basic file read with path argument",
-    ),
-    ToolCallTest(
-        id="read_file_offset",
-        name="Read File — with offset",
-        prompt="Read lines 50 through 80 of /var/log/system.log",
-        expected_tool="read_file",
-        expected_args={"path": "/var/log/system.log"},
-        description="File read with offset parameter",
-    ),
-    ToolCallTest(
-        id="web_search_basic",
-        name="Web Search — basic query",
-        prompt="Search the web for 'TurboQuant KV cache compression benchmarks'",
-        expected_tool="web_search",
-        expected_args={"query": "turboquant"},
-        description="Web search with query containing keywords",
-    ),
-    ToolCallTest(
-        id="terminal_basic",
-        name="Terminal — simple command",
-        prompt="Run `ls -la /tmp` to see what files are there.",
-        expected_tool="terminal",
-        expected_args={"command": "ls"},
-        description="Terminal command execution",
-    ),
-    ToolCallTest(
-        id="terminal_complex",
-        name="Terminal — complex command",
-        prompt="Check the disk usage of the current directory with `du -sh .`",
-        expected_tool="terminal",
-        expected_args={"command": "du"},
-        description="Terminal with different command",
-    ),
-    ToolCallTest(
-        id="code_exec_basic",
-        name="Code Execution — python",
-        prompt="Run this Python code: print(sum(range(100)))",
-        expected_tool="execute_code",
-        expected_args={"code": "sum"},
-        description="Code execution with Python",
-    ),
-    ToolCallTest(
-        id="code_exec_complex",
-        name="Code Execution — multi-line",
-        prompt="Write and run Python code that reads a CSV file and counts the rows. Use the csv module.",
-        expected_tool="execute_code",
-        expected_args={"code": "csv"},
-        description="Code execution with multi-line Python",
-    ),
-    ToolCallTest(
-        id="delegate_basic",
-        name="Delegate Task — simple",
-        prompt="Delegate this task to a subagent: research the latest llama.cpp release notes.",
-        expected_tool="delegate_task",
-        expected_args={"goal": "llama"},
-        description="Task delegation with goal",
-    ),
-    ToolCallTest(
-        id="delegate_context",
-        name="Delegate Task — with context",
-        prompt="Spawn a subagent to review the Python files in /src. Context: look for security issues.",
-        expected_tool="delegate_task",
-        expected_args={"goal": "review"},
-        description="Task delegation with context",
-    ),
-    ToolCallTest(
-        id="parallel_two",
-        name="Parallel Tools — two in one response",
-        prompt="Read the file /etc/hostname AND check the current date by running `date`. Do both at the same time.",
-        expected_tool="read_file",  # at least one of the two
-        expected_args={"path": "/etc/hostname"},
-        description="Two tool calls in a single response",
-        # Note: this test checks that at least 2 tool calls are returned
-    ),
-]
-
-
-# ── Result types ──────────────────────────────────────────────────
-
-@dataclass
-class TestResult:
-    id: str
-    name: str
-    passed: bool
-    tool_called: Optional[str] = None
-    expected_tool: str = ""
-    schema_valid: bool = False
-    args_valid: bool = False
-    latency_ms: float = 0.0
-    raw_response: str = ""
-    error: Optional[str] = None
-
-
-@dataclass
-class SuiteResult:
-    timestamp: str
-    model: str
-    endpoint: str
-    kv_type: str
-    total: int = 0
-    passed: int = 0
-    failed: int = 0
-    accuracy: float = 0.0
-    meets_threshold: bool = False
-    threshold: float = 0.95
-    results: list = field(default_factory=list)
-    error: Optional[str] = None
-
-
-# ── Schema validation ────────────────────────────────────────────
-
-def validate_tool_call_schema(call: dict) -> bool:
-    """Validate that a tool call response has the expected structure."""
-    if not isinstance(call, dict):
-        return False
-
-    # OpenAI format: { "type": "function", "function": { "name": "...", "arguments": "{}" } }
-    if call.get("type") == "function":
-        func = call.get("function", {})
-        return (
-            isinstance(func.get("name"), str) and len(func["name"]) > 0
-            and isinstance(func.get("arguments"), str)
-        )
-
-    # Alternative format: { "name": "...", "arguments": "{}" }
-    if "name" in call and "arguments" in call:
-        return (
-            isinstance(call["name"], str) and len(call["name"]) > 0
-            and isinstance(call["arguments"], str)
-        )
-
-    return False
-
-
-def validate_tool_args(args_str: str, expected: dict) -> bool:
-    """Validate that tool arguments contain expected keys/values."""
-    try:
-        args = json.loads(args_str)
-    except (json.JSONDecodeError, TypeError):
-        return False
-
-    if not isinstance(args, dict):
-        return False
-
-    for key, value in expected.items():
-        if key not in args:
-            return False
-        # For string values, check substring match
-        if isinstance(value, str) and isinstance(args[key], str):
-            if value.lower() not in args[key].lower():
-                return False
-        # For non-string values, check exact match
-        elif args[key] != value:
-            return False
-
-    return True
-
-
-def extract_tool_calls(response: dict) -> list:
-    """Extract tool calls from an API response."""
-    choices = response.get("choices", [])
-    if not choices:
-        return []
-
-    message = choices[0].get("message", {})
-
-    # Standard OpenAI format
-    tool_calls = message.get("tool_calls", [])
-    if tool_calls:
-        return tool_calls
-
-    # Some models return tool calls in content as JSON
-    content = message.get("content", "")
-    if content:
-        # Try to parse content as JSON tool call
-        try:
-            parsed = json.loads(content)
-            if isinstance(parsed, dict) and "name" in parsed and "arguments" in parsed:
-                return [parsed]
-            if isinstance(parsed, list):
-                return [c for c in parsed if isinstance(c, dict) and "name" in c]
-        except (json.JSONDecodeError, TypeError):
-            # Look for JSON blocks in content
-            json_match = re.search(r'\{[^{}]*"name"\s*:\s*"[^"]*"[^{}]*\}', content)
-            if json_match:
-                try:
-                    return [json.loads(json_match.group())]
-                except json.JSONDecodeError:
-                    pass
-
-    return []
-
-
-# ── API interaction ───────────────────────────────────────────────
-
-def call_model(endpoint: str, model: str, messages: list, tools: list,
-               temperature: float = 0.1, timeout: int = 60) -> dict:
-    """Call the model via OpenAI-compatible API."""
-    import urllib.request
-
-    payload = json.dumps({
-        "model": model,
-        "messages": messages,
-        "tools": tools,
-        "temperature": temperature,
-        "max_tokens": 1024,
-    }).encode()
-
-    req = urllib.request.Request(
-        f"{endpoint}/chat/completions",
-        data=payload,
-        headers={"Content-Type": "application/json"},
-        method="POST",
-    )
-
-    start = time.time()
-    try:
-        resp = urllib.request.urlopen(req, timeout=timeout)
-        data = json.loads(resp.read())
-        data["_latency_ms"] = (time.time() - start) * 1000
-        return data
-    except Exception as e:
-        return {"error": str(e), "_latency_ms": (time.time() - start) * 1000}
-
-
-# ── Test runner ───────────────────────────────────────────────────
-
-def run_single_test(endpoint: str, model: str, test: ToolCallTest) -> TestResult:
-    """Run a single tool call test."""
-    messages = [
-        {
-            "role": "system",
-            "content": (
-                "You are a helpful assistant. When the user asks you to perform "
-                "a task, use the appropriate tool. Always call exactly one tool "
-                "unless the user explicitly asks for multiple things."
-            ),
-        },
-        {"role": "user", "content": test.prompt},
-    ]
-
-    response = call_model(endpoint, model, messages, TOOL_SCHEMAS)
-
-    if "error" in response:
-        return TestResult(
-            id=test.id,
-            name=test.name,
-            passed=False,
-            expected_tool=test.expected_tool,
-            error=response["error"],
-            latency_ms=response.get("_latency_ms", 0),
-        )
-
-    tool_calls = extract_tool_calls(response)
-    latency = response.get("_latency_ms", 0)
-
-    if not tool_calls:
-        # Model didn't call any tool
-        content = response.get("choices", [{}])[0].get("message", {}).get("content", "")
-        return TestResult(
-            id=test.id,
-            name=test.name,
-            passed=False,
-            expected_tool=test.expected_tool,
-            latency_ms=latency,
-            raw_response=content[:500],
-            error="No tool call returned",
-        )
-
-    # Validate first tool call
-    call = tool_calls[0]
-    schema_valid = validate_tool_call_schema(call)
-
-    # Extract tool name
-    if call.get("type") == "function":
-        tool_name = call["function"]["name"]
-        args_str = call["function"]["arguments"]
-    else:
-        tool_name = call.get("name", "")
-        args_str = call.get("arguments", "{}")
-
-    args_valid = validate_tool_args(args_str, test.expected_args)
-    tool_correct = tool_name == test.expected_tool
-    passed = tool_correct and schema_valid
-
-    return TestResult(
-        id=test.id,
-        name=test.name,
-        passed=passed,
-        tool_called=tool_name,
-        expected_tool=test.expected_tool,
-        schema_valid=schema_valid,
-        args_valid=args_valid,
-        latency_ms=latency,
-        raw_response=json.dumps(tool_calls[:2])[:500],
-    )
-
-
-def run_dry_run() -> SuiteResult:
-    """Validate schemas and test structure without a running server."""
-    print("=== DRY RUN — Schema Validation Only ===\n")
-
-    results = []
-    for test in TEST_CASES:
-        # Validate schemas parse
-        schema_valid = True
-        for tool in TOOL_SCHEMAS:
-            try:
-                assert "type" in tool
-                assert tool["type"] == "function"
-                func = tool["function"]
-                assert "name" in func
-                assert "description" in func
-                assert "parameters" in func
-                params = func["parameters"]
-                assert "type" in params
-                assert "properties" in params
-            except AssertionError:
-                schema_valid = False
-
-        results.append(TestResult(
-            id=test.id,
-            name=test.name,
-            passed=schema_valid,
-            expected_tool=test.expected_tool,
-            schema_valid=schema_valid,
-            args_valid=True,
-        ))
-
-    passed = sum(1 for r in results if r.passed)
-    suite = SuiteResult(
-        timestamp=datetime.now(timezone.utc).isoformat(),
-        model="dry-run",
-        endpoint="none",
-        kv_type="none",
-        total=len(results),
-        passed=passed,
-        failed=len(results) - passed,
-        accuracy=passed / len(results) if results else 0,
-        meets_threshold=passed == len(results),
-        threshold=1.0,
-        results=[asdict(r) for r in results],
-    )
-
-    return suite
-
-
-def run_suite(endpoint: str, model: str, kv_type: str, runs: int = 1,
-              threshold: float = 0.95) -> SuiteResult:
-    """Run the full tool call regression suite."""
-    print(f"=== TurboQuant Tool Call Regression Suite ===")
-    print(f"Endpoint: {endpoint}")
-    print(f"Model:    {model}")
-    print(f"KV Type:  {kv_type}")
-    print(f"Runs:     {runs}")
-    print(f"Threshold: {threshold:.0%}")
-    print()
-
-    # Check server is reachable
-    try:
-        import urllib.request
-        health_req = urllib.request.Request(f"{endpoint}/models", method="GET")
-        urllib.request.urlopen(health_req, timeout=5)
-    except Exception as e:
-        return SuiteResult(
-            timestamp=datetime.now(timezone.utc).isoformat(),
-            model=model,
-            endpoint=endpoint,
-            kv_type=kv_type,
-            error=f"Server unreachable: {e}",
-        )
-
-    all_results = []
-    for run_idx in range(runs):
-        if runs > 1:
-            print(f"\n--- Run {run_idx + 1}/{runs} ---")
-
-        for test in TEST_CASES:
-            print(f"  {test.id}: ", end="", flush=True)
-            result = run_single_test(endpoint, model, test)
-            status = "PASS" if result.passed else "FAIL"
-            tool_info = f"called={result.tool_called}" if result.tool_called else "no tool"
-            print(f"{status} ({tool_info}, {result.latency_ms:.0f}ms)")
-            if result.error:
-                print(f"         Error: {result.error}")
-            all_results.append(result)
-
-    passed = sum(1 for r in all_results if r.passed)
-    total = len(all_results)
-    accuracy = passed / total if total > 0 else 0
-
-    suite = SuiteResult(
-        timestamp=datetime.now(timezone.utc).isoformat(),
-        model=model,
-        endpoint=endpoint,
-        kv_type=kv_type,
-        total=total,
-        passed=passed,
-        failed=total - passed,
-        accuracy=accuracy,
-        meets_threshold=accuracy >= threshold,
-        threshold=threshold,
-        results=[asdict(r) for r in all_results],
-    )
-
-    print(f"\n{'='*60}")
-    print(f"RESULTS: {passed}/{total} passed ({accuracy:.1%})")
-    print(f"Threshold: {threshold:.0%}")
-    print(f"VERDICT: {'PASS' if suite.meets_threshold else 'FAIL'}")
-    print(f"{'='*60}")
-
-    return suite
-
-
-# ── Markdown report ───────────────────────────────────────────────
-
-def generate_report(suite: SuiteResult, output_path: str) -> None:
-    """Generate a markdown results matrix."""
-    lines = [
-        "# Tool Call Regression Results",
-        "",
-        f"**Generated:** {suite.timestamp}",
-        f"**Model:** {suite.model}",
-        f"**Endpoint:** {suite.endpoint}",
-        f"**KV Type:** {suite.kv_type}",
-        "",
-        "## Summary",
-        "",
-        f"| Metric | Value |",
-        f"|--------|-------|",
-        f"| Total tests | {suite.total} |",
-        f"| Passed | {suite.passed} |",
-        f"| Failed | {suite.failed} |",
-        f"| Accuracy | {suite.accuracy:.1%} |",
-        f"| Threshold | {suite.threshold:.0%} |",
-        f"| Verdict | {'PASS' if suite.meets_threshold else 'FAIL'} |",
-        "",
-        "## Test Matrix",
-        "",
-        "| Test ID | Tool Expected | Tool Called | Schema | Args | Latency | Status |",
-        "|---------|--------------|-------------|--------|------|---------|--------|",
-    ]
-
-    for r in suite.results:
-        d = r if isinstance(r, dict) else asdict(r)
-        status = "PASS" if d["passed"] else "FAIL"
-        schema = "OK" if d.get("schema_valid") else "FAIL"
-        args = "OK" if d.get("args_valid") else "FAIL"
-        called = d.get("tool_called") or "none"
-        latency = f"{d.get('latency_ms', 0):.0f}ms"
-        lines.append(
-            f"| {d['id']} | {d['expected_tool']} | {called} | {schema} | {args} | {latency} | {status} |"
-        )
-
-    if suite.error:
-        lines.extend(["", "## Error", "", suite.error])
-
-    os.makedirs(os.path.dirname(output_path), exist_ok=True)
-    with open(output_path, "w") as f:
-        f.write("\n".join(lines) + "\n")
-    print(f"\nReport saved to {output_path}")
-
-
-# ── Main ──────────────────────────────────────────────────────────
-
-def main():
-    parser = argparse.ArgumentParser(description="TurboQuant Tool Call Regression Suite")
-    parser.add_argument("--endpoint", default="http://localhost:8081/v1",
-                        help="llama.cpp OpenAI-compatible endpoint")
-    parser.add_argument("--model", default="gemma-4", help="Model name")
-    parser.add_argument("--kv-type", default="turbo4", help="KV cache type being tested")
-    parser.add_argument("--runs", type=int, default=1, help="Number of runs per test")
-    parser.add_argument("--threshold", type=float, default=0.95,
-                        help="Minimum accuracy to pass (0.0-1.0)")
-    parser.add_argument("--output", default="benchmarks/tool-call-regression.md",
-                        help="Output markdown report path")
-    parser.add_argument("--results-json", default="benchmarks/tool-call-regression.json",
-                        help="Output JSON results path")
-    parser.add_argument("--dry-run", action="store_true",
-                        help="Validate schemas only, no server needed")
-    args = parser.parse_args()
-
-    if args.dry_run:
-        suite = run_dry_run()
-    else:
-        suite = run_suite(
-            endpoint=args.endpoint,
-            model=args.model,
-            kv_type=args.kv_type,
-            runs=args.runs,
-            threshold=args.threshold,
-        )
-
-    # Save results
-    generate_report(suite, args.output)
-    os.makedirs(os.path.dirname(args.results_json), exist_ok=True)
-    with open(args.results_json, "w") as f:
-        json.dump(asdict(suite), f, indent=2)
-    print(f"JSON results saved to {args.results_json}")
-
-    # Exit code: 0 if passes threshold, 1 otherwise
-    sys.exit(0 if suite.meets_threshold else 1)
-
-
-if __name__ == "__main__":
-    main()