{"id":"mellum2-1","slug":"mellum2-1","name":"Mellum2.1","orgId":"jetbrains","family":"mellum2-1","releaseDate":null,"availability":"ga","openWeights":true,"license":"Apache 2.0, per the Hugging Face card metadata and text and the JetBrains blog; the Hugging Face repo has no separate LICENSE file","paramsTotal":12149923072,"paramsSource":"tracked","contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://blog.jetbrains.com/ai/2026/10/mellum2-1-gets-to-work-a-fast-open-model-for-coding-agents/","hfRepo":null,"modalities":["text"],"aliases":["Mellum2.1","mellum2.1","mellum 2.1","mellum2.1 thinking","mellum2.1-12b-a2.5b-thinking","jetbrains/mellum2.1-12b-a2.5b-thinking","jetbrains/mellum2.1-12b-a2.5b-thinking-gguf","jetbrains/mellum2.1-12b-a2.5b-thinking-mtp"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-11-01","announced":"2026-10-08","blurb":"JetBrains' open-weights coding model for agents and sub-agents: a 12B mixture-of-experts thinking model with 2.5B parameters active per token and a 131,072-token context. It keeps the Mellum2 architecture unchanged; the update comes from reinforcement-learning post-training in sandboxed coding environments. Weights on Hugging Face under Apache 2.0, with GGUF builds and a multi-token-prediction head for faster vLLM serving. No hosted API is stated; it is meant to run on your own hardware.","api":"none","openWeights":"yes","params":{"text":"About 12.15B total (12,149,923,072 BF16 parameters), 2.5B active per token","source":"Hugging Face safetensors metadata and model card for JetBrains/Mellum2.1-12B-A2.5B-Thinking","total":12149923072},"facts":[{"label":"Type","value":"Text-only thinking model (reasons in <think> blocks) for coding agents: exploring a repository, editing files and checking its own work"},{"label":"Architecture","value":"Mixture of experts, 12B total and 2.5B active: 64 experts with 8 active per token; unchanged from Mellum2"},{"label":"Context","value":"131,072 tokens"},{"label":"Base model","value":"JetBrains/Mellum2-12B-A2.5B-Base (per the Hugging Face card); the next version of Mellum2 Thinking, not a fine-tune of it"},{"label":"Training","value":"Post-training mostly reinforcement learning, with millions of sandboxed runs across thousands of environments, per JetBrains"},{"label":"Weights","value":"Hugging Face JetBrains/Mellum2.1-12B-A2.5B-Thinking (BF16, not gated)"},{"label":"Variants","value":"JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF (BF16, Q8_0, Q6_K, Q4_K_M and MXFP4_MOE quantizations plus MTP head files) and JetBrains/Mellum2.1-12B-A2.5B-Thinking-MTP (the same weights plus a multi-token-prediction head for speculative decoding in vLLM, added Oct 30, 2026)"},{"label":"License","value":"Apache 2.0, per the Hugging Face card metadata and text and the JetBrains blog; the Hugging Face repo has no separate LICENSE file"},{"label":"API","value":"None stated; self-hosted (vLLM, llama.cpp, Ollama, LM Studio)"},{"label":"Related report","value":"Mellum2 technical report, arXiv 2605.31268"}],"claims":["Lab claim: in JetBrains' own thinking-mode evaluation, 82.0 on LiveCodeBench v6, 91.5 on HumanEval+, 79.4 on MBPP+, 83.3 on AIME 25/26 and 64.6 on GPQA Diamond.","Lab claim: 47.0 on SWE-bench Verified, 28.0 on SWE-bench Pro and 17.4 on Terminal-Bench 2.1 with the Pi v0.73.1 agent harness, against 2.0, 0.0 and 0.6 for Mellum2 Thinking and 50.0, 38.0 and 21.7 for Qwen3.5 (9B).","Lab claim: 62.3 on BFCL v4, 90.6 on IFEval and 87.8 on MMLU-Redux, with an 8.5% HarmBench harmful rate.","Lab claim: under heavy load it serves almost twice as many tokens as Qwen3.5-9B, and MTP makes a single request about 1.6 times faster."],"sources":[{"kind":"lab","title":"Mellum2.1 Gets to Work: A Fast Open Model for Coding Agents (JetBrains AI blog)","url":"https://blog.jetbrains.com/ai/2026/10/mellum2-1-gets-to-work-a-fast-open-model-for-coding-agents/","date":"2026-10-08"},{"kind":"repo","title":"JetBrains/Mellum2.1-12B-A2.5B-Thinking on Hugging Face","url":"https://huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking"},{"kind":"repo","title":"JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF on Hugging Face","url":"https://huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF"},{"kind":"repo","title":"JetBrains/Mellum2.1-12B-A2.5B-Thinking-MTP on Hugging Face","url":"https://huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking-MTP"},{"kind":"paper","title":"Mellum2 technical report (arXiv)","url":"https://arxiv.org/abs/2605.31268"}],"firstBenchmark":null},"organization":{"id":"jetbrains","name":"JetBrains","slug":"jetbrains","country":null,"url":"https://www.jetbrains.com","color":"#11110F","mark":"J"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Oct 8, 2026; no independent results yet.","lead":"As of Nov 22, 2026, Mellum2.1 is pending benchmarks on isitsota: announced Oct 8, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}