gent (Colle et al., 2025)
29,529
GSM8K (TIR) (Cobbe et al., 2021)
7,473
Toucan 1.5M
117,586
EnvScaler SFT Trajectories
9,022
OpenSeeker v1
4,949
Subtotal
733,714
Multilingual
SmolTalk2 Multilingual (8 lang) (Bakouch et al., 2025)
653,834
25.83%
EuroBlocks Multilingual (2512) (Ramos et al., 2026)
286,381
EuroBlocks Aya (Synthetic 1124) (Martins et al., 2025)
17,467
Subtotal
957,682
Regional
WikiQA (Yang et al., 2015)
482,799
15.35%
Romansh Varieties (Hernández-Cano et al., 2026)
46,078
German→\rightarrowRomansh MT
30,000
African Languages (Hernández-Cano et al., 2026)
7,339
Swiss German Dialects (Hernández-Cano et al., 2026)
2,674
Swiss Charter Q&A (Hernández-Cano et al., 2026)
88
Subtotal
568,978
Domain-Specific
MedMCQA (Pal et al., 2022)
156,355
7.26%
MIRIAD (Zheng et al., 2025)
99,556
Medical-o1 (Chen et al., 2025b)
13,383
Subtotal
269,294
Total
3,707,851
100.00%
Tool-Use Data.
We select tool-use datasets by whether their trajectories are execution-verified, rather than filtered only by an LLM judge, and drop candidates that do not meet this bar. The mixture combines a curated subset of Toucan (Xu et al., 2026) (119,287 samples), EnvScaler (Song et al., 2026) (9,022), OpenSeeker (Du et al., 2026) (4,949), an execution-verified subset of the Jupyter Agent dataset (Colle et al., 2025) (roughly 30,000 sandbox-verified notebooks), Glaive function calling (Glaive AI, 2023) (roughly 52,000), and GSM8K (Cobbe et al., 2021) trajectories with calculator tool use (roughly 7,400). We validate and clean the tool-use datasets to ensure correct conversation formatting, declared tools, valid call arguments, and matching nonempty tool outputs. Because a model that sees tools only in tool-calling samples never learns to ignore them, we additionally inject tool declarations into 131,557 conversations that do not call any tool, teaching the model to answer directly even when tools are available. A cleanup pass strips 950,544 calculator-annotation blocks that appear in samples without a corresponding tool declaration.
Multilingual and Safety Data.
We use updated versions of EuroBlocks-SFT-2512 (Ramos et al., 2026) and Llama-Nemotron Safety (Bercovich et al., 2025) to improve our safety and multilingual capabilities. For Llama-Nemotron Safety, we observe over-refusals of benign prompts and, on some occasions, harmful responses to dangerous prompts; therefore, we regenerate responses with GPT-OSS-120B-Safeguard and verify their validity with a Qwen3-235B judge to improve assistant behavior and mitigate over-refusal. In the EuroBlocks-SFT-2512 dataset, we observe that some of the assistant responses contain traces of the original EuroLLM identity training with the model name present in the provided outputs. Since we cannot remove something like that verbatim, we use a pool of several models (GPT-OSS-120B, MiniMax-M2, Qwen3.5-397B) randomly sampled to regenerate responses with the Apertus identity provided in the system prompt. Of the 1M original samples, we use only 286k non-English conversations to diversify our multilingual training mix.
Decontamination.
We decontaminate the SFT mixture against a list of 595 evaluation benchmarks with n-gram matching, and manually verify the flagged samples of every affected dataset rather than dropping hits blindly, since thresholds tuned for English text over-trigger on multimodal boilerplate. The true positives concentrate in multilingual datasets that contain translated benchmark items: in PangeaInstruct, roughly 71% of the 9,114 flagged samples are genuine benchmark translations, estimated from a manual inspection of 100 random hits. For EuroBlocks 2512, almost all of the removed samples are unrelated non-contaminated multilingual samples. For our instruction-following mix, the majority of “flagged” samples contain an instruction format similar to some from the IFEval benchmark, which does not necessarily mean direct contamination. We still decide to keep the original similarity threshold of 0.5 (as used in