90 lines
4.5 KiB
Markdown
90 lines
4.5 KiB
Markdown
# Triage Report — AI model releases & agent funding (Oct 13–28, 2026)
|
||
|
||
Date: 2026-10-28
|
||
Sources: ArtificialWatch (artificialwatch.com), BenchLM (benchlm.ai), Bloomberg, Fortune,
|
||
TechCrunch, TechFundingNews, SecurityWeek, Sifted.
|
||
|
||
## Summary
|
||
- Items reviewed: 15
|
||
- P0 action-now: 5
|
||
- P1 next-up: 6
|
||
- P2 monitor: 4
|
||
- BACKLOG: 0
|
||
|
||
## P0 — ACTION NOW (market-moving / affects our stack)
|
||
|
||
1. [P0] Nous Research — Raised $90M Series B at a $1.5B valuation ($158M total); Hermes
|
||
Agent cloned 24M+ times, ~2.5% of global AI token usage (TechCrunch, 2026-10-07).
|
||
Next: note the open-core pressure on our agent usage share and watch for 'Hermes for
|
||
Businesses' enterprise product.
|
||
Source: https://techcrunch.com/2026/10/07/nous-research-confirms-it-hit-1-5b-valuation-launches-ai-agents-for-business-users/
|
||
|
||
2. [P0] Daytona — $60M Series B at a $1B valuation ($91M total); round led by FirstMark
|
||
with Stripe (Collison), Shopify (Lütke), Perplexity (Srinivas), Zoom (Yuan) angel
|
||
investing; Qualcomm Ventures, Samsung Next, Okta Ventures, Evergreen Fund, Alumni
|
||
Ventures, Torch Capital (Fortune, 2026-10-20).
|
||
Next: product watch — agent sandbox cloud vs Modal (raised $750M at $15.75B).
|
||
Source: https://fortune.com/2026/10/20/daytona-startup-building-the-agentic-cloud-valued-at-1-billion-in-latest-funding-round/
|
||
|
||
3. [P0] Hyphenate — $45M Series A led by Nexus Venture Partners (total $54M); revenue
|
||
up 86x YoY; agents cover order-to-cash, treasury, GL accounting, procure-to-pay,
|
||
reporting & insights; integrates with ERPs and banks (TechFundingNews, 2026-10-27).
|
||
Source: https://techfundingnews.com/hyphenate-45m-series-a-nexus-b-capital-ai-finance-agents/
|
||
|
||
4. [P0] Mave — $28M Series A (total ~$34M) by Unit 8200 alumni (Guy Eshel CTO, Raziel
|
||
Oren CEO); SecOps AI agents investigating alerts across security stacks; ~30 staff
|
||
growing (SecurityWeek, 2026-10-25).
|
||
Source: https://www.securityweek.com/former-unit-8200-officers-raise-34-million-for-agentic-secops-startup-mave/
|
||
|
||
5. [P0] Rein Security — $25M Series A (total $35M); co-led by Glilot Capital and Sienna
|
||
Venture Capital; runtime security + real-time context for AI agents; Tel Aviv/NYC
|
||
(SecurityWeek, 2026-10-08).
|
||
Source: https://www.securityweek.com/rein-security-raises-25-million-to-guard-ai-agents-at-runtime/
|
||
|
||
## P1 — NEXT UP (watch this batch)
|
||
|
||
6. [P1] Flint — Recursive, launched 2026-10-21; leads IFBench Oct 2026 at 89.2% vs
|
||
MAI-Thinking-1 85% and Qwen3.8 Max 82.8% (BenchLM). Source:
|
||
https://benchlm.ai/benchmarks/ifbench
|
||
|
||
7. [P1] Gemini 4 — Google, frontier, launched 2026-10-13, $2/$10 per 1M tokens,
|
||
bench index 80 (ArtificialWatch). Source:
|
||
https://artificialwatch.com/releases
|
||
|
||
8. [P1] Kimi K3.1 Pro — Moonshot, frontier, launched 2026-10-23, $2/$10 per 1M tokens,
|
||
bench index 71 (ArtificialWatch). Source:
|
||
https://artificialwatch.com/releases
|
||
|
||
9. [P1] GPT-6.1 Terra — OpenAI, launched 2026-10-26, $1/$5 per 1M tokens, bench index
|
||
75; topped BenchLM's Oct 2026 race at 76.06/100 (ArtificialWatch + BenchLM). Source:
|
||
https://benchlm.ai/ai-race
|
||
|
||
10. [P1] Gemini 4 Carbon — Google, confirmed 2026-10-28 (exact day), benchmarks tracked
|
||
(BenchLM). Source: https://benchlm.ai/model-updates/releases/october-2026
|
||
|
||
11. [P1] LFM3 — LiquidAI, 7 model variants, launched 2026-10-27 via Hugging Face
|
||
(BenchLM). Source: https://benchlm.ai/model-updates/releases/october-2026
|
||
|
||
## P2 — MONITOR (minor releases, no immediate action)
|
||
|
||
12. [P2] Mercury 3 — Inception, launched 2026-10-19 (dated launch post). Source:
|
||
https://benchlm.ai/model-updates/releases/october-2026
|
||
|
||
13. [P2] MiniMax M3.1 Flash — MiniMax, launched 2026-10-19, $0.15/$0.60 per 1M tokens,
|
||
bench index 36.8 (ArtificialWatch). Source:
|
||
https://artificialwatch.com/releases
|
||
|
||
14. [P2] Voxtral 3 — Mistral AI, 4 model variants, launched 2026-10-20. Source:
|
||
https://benchlm.ai/model-updates/releases/october-2026
|
||
|
||
15. [P2] Nemotron 3.5 Bolt E4B 4GB — NVIDIA, 2 model variants, launched 2026-10-23
|
||
(4GB GGUF checkpoint; evidence: NVIDIA research page). Source:
|
||
https://benchlm.ai/model-updates/releases/october-2026
|
||
|
||
## Method notes
|
||
- Per-item dates are the day each release was verified live/public, not the announcement
|
||
date (ArtificialWatch practice); BenchLM practice likewise.
|
||
- "17 models" (ArtificialWatch) vs "44/45 releases" (BenchLM) are different trackers;
|
||
do not average them. Per-item sources above are authoritative for each row.
|
||
- Unverified claims are labeled as claims of the cited source.
|