{ "hf_id": "JetBrains/Mellum2-12B-A2.5B-Instruct", "meta": { "title": "Mellum2-12B-A2.5B-Instruct", "slug": "mellum2-12b-a2.5b-instruct", "provider": "JetBrains", "description": "JetBrains' instruction-tuned code MoE (12B total / 2.5B active) that answers directly without an externalized chain of thought — low-latency coding and tool use", "date_added": "2026-06-02", "date_updated": "2026-06-02", "tasks": [ "text" ], "performance_headline": "78.4 EvalPlus, 67.1 MultiPL-E — direct answers, fits on a single GPU", "related_recipes": [ "JetBrains/Mellum2-12B-A2.5B-Thinking" ], "hardware": { "h200": "verified", "mi300x": "verified", "mi325x": "verified", "mi355x": "verified" } }, "recommended_command": { "hardware": "h200", "strategy": "single_node_tp", "variant": "default", "node_count": 1, "deploy_type": "single_node", "env": {}, "docker_image": "vllm/vllm-openai:latest", "command": "vllm serve JetBrains/Mellum2-12B-A2.5B-Instruct \\\n --max-model-len 131072 \\\n --trust-remote-code \\\n --tensor-parallel-size 1 \\\n --enable-auto-tool-choice \\\n --tool-call-parser hermes", "argv": [ "vllm", "serve", "JetBrains/Mellum2-12B-A2.5B-Instruct", "--max-model-len", "131072", "--trust-remote-code", "--tensor-parallel-size", "1", "--enable-auto-tool-choice", "--tool-call-parser", "hermes" ], "docker_command": "docker run --gpus all \\\n --privileged --ipc=host -p 8000:8000 \\\n -v ~/.cache/huggingface:/root/.cache/huggingface \\\n vllm/vllm-openai:latest JetBrains/Mellum2-12B-A2.5B-Instruct \\\n --max-model-len 131072 \\\n --trust-remote-code \\\n --tensor-parallel-size 1 \\\n --enable-auto-tool-choice \\\n --tool-call-parser hermes", "docker_argv": [ "docker", "run", "--gpus", "all", "--privileged", "--ipc=host", "-p", "8000:8000", "-v", "~/.cache/huggingface:/root/.cache/huggingface", "vllm/vllm-openai:latest", "JetBrains/Mellum2-12B-A2.5B-Instruct", "--max-model-len", "131072", "--trust-remote-code", "--tensor-parallel-size", "1", "--enable-auto-tool-choice", "--tool-call-parser", "hermes" ], "strategy_spec": { "name": "single_node_tp", "deploy_type": "single_node", "display_name": "Tensor Parallel", "short_label": "TP", "orientation": "latency", "description": "Single-node tensor parallel. Splits the model across all local GPUs. TP size is set to the GPU count at deploy time. The simplest multi-GPU strategy — works for all model architectures.\n", "hardware_match": { "min_gpus": 1, "max_gpus": 8, "multi_node": false }, "vllm_args": [], "parallel_flag": "--tensor-parallel-size" }, "hardware_profile": { "brand": "NVIDIA", "generation": "hopper", "compute_arch": "sm90", "display_name": "H200", "description": "NVIDIA H200 SXM 141 GB HBM3e · 8-GPU HGX node", "gpu_count": 8, "vram_gb": 1128, "multi_node": false }, "alternatives": { "single_node_tep": "/JetBrains/Mellum2-12B-A2.5B-Instruct/strategies/single_node_tep.json", "single_node_dep": "/JetBrains/Mellum2-12B-A2.5B-Instruct/strategies/single_node_dep.json", "kv_store_centralized_mooncake": "/JetBrains/Mellum2-12B-A2.5B-Instruct/strategies/kv_store_centralized_mooncake.json", "kv_store_distributed_mooncake": "/JetBrains/Mellum2-12B-A2.5B-Instruct/strategies/kv_store_distributed_mooncake.json" }, "by_hardware": { "h200": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/h200.json", "h100": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/h100.json", "b200": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/b200.json", "gb200": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/gb200.json", "b300": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/b300.json", "gb300": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/gb300.json", "mi300x": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/mi300x.json", "mi325x": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/mi325x.json", "mi350x": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/mi350x.json", "mi355x": "/JetBrains/Mellum2-12B-A2.5B-Instruct/hw/mi355x.json" } }, "model": { "model_id": "JetBrains/Mellum2-12B-A2.5B-Instruct", "min_vllm_version": "0.23.0", "architecture": "moe", "parameter_count": "12B", "active_parameters": "2.5B", "context_length": 131072, "base_args": [ "--max-model-len", "131072", "--trust-remote-code" ], "base_env": {}, "install": { "pip": { "command": "uv venv\nsource .venv/bin/activate\nuv pip install -U vllm --torch-backend auto" }, "docker": { "command": "docker pull vllm/vllm-openai:latest" } } }, "features": { "tool_calling": { "description": "Hermes tool-call parser for function calling", "args": [ "--enable-auto-tool-choice", "--tool-call-parser", "hermes" ] } }, "opt_in_features": [], "variants": { "default": { "precision": "bf16", "vram_minimum_gb": 29, "description": "Native bfloat16 weights; fits comfortably on a single H200/H100/A100" } }, "hardware_overrides": {}, "guide": "## Overview\n\n[Mellum2-12B-A2.5B-Instruct](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Instruct)\nis JetBrains' instruction-tuned code assistant. It shares the same Mixture-of-Experts\nbackbone as the rest of the Mellum2 family — 64 experts (8 activated per token), 12B total\n/ 2.5B active parameters, sliding-window + full-attention layers, 131,072-token context —\nbut is post-trained (SFT + RLVR on math, coding, tool use, instruction following, reasoning,\nand knowledge) to **answer directly, without an externalized chain of thought**. For complex\ndebugging, multi-step planning, or math/reasoning-heavy tasks where you want explicit\nreasoning traces, use the\n[Thinking](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Thinking) variant instead.\n\n## Prerequisites\n\n- Hardware: a single H200, H100, or A100 (~29 GB at bf16) is plenty. On AMD,\n use MI300X / MI325X / MI355X with TP=1.\n- vLLM **nightly** — `MellumForCausalLM` support landed after v0.22.0 and is not yet in a\n stable release. Install the nightly wheels until the next tagged release ships.\n\n### Install vLLM (nightly)\n\n```bash\nuv venv\nsource .venv/bin/activate\nuv pip install -U vllm --extra-index-url https://wheels.vllm.ai/nightly\n```\n\n## Launch command\n\nUnlike the Thinking checkpoint, Instruct does not emit `` blocks, so no\n`--reasoning-parser` is needed.\n\n### Single GPU\n\n```bash\nvllm serve JetBrains/Mellum2-12B-A2.5B-Instruct \\\n --max-model-len 131072 \\\n --trust-remote-code \\\n --tensor-parallel-size 1\n\n# Optional: add tool calling.\nvllm serve JetBrains/Mellum2-12B-A2.5B-Instruct \\\n --max-model-len 131072 \\\n --trust-remote-code \\\n --tensor-parallel-size 1 \\\n --enable-auto-tool-choice \\\n --tool-call-parser hermes\n```\n\n### Docker (AMD MI300X / MI325X / MI355X)\n\nMI300X / MI325X / MI355X GPUs have larger per-GPU HBM, so you can use\n`--max-model-len auto`.\n\n```bash\n# Plain serving\nvllm serve JetBrains/Mellum2-12B-A2.5B-Instruct \\\n --max-model-len auto \\\n --trust-remote-code \\\n --tensor-parallel-size 1\n\n# Add tool calling.\nvllm serve JetBrains/Mellum2-12B-A2.5B-Instruct \\\n --max-model-len auto \\\n --trust-remote-code \\\n --tensor-parallel-size 1 \\\n --enable-auto-tool-choice \\\n --tool-call-parser hermes\n```\n\n## Client usage\n\nJetBrains recommends sampling at `temperature=0.6`, `top_p=0.95`, `top_k=20`.\n\n```python\nfrom openai import OpenAI\nclient = OpenAI(base_url=\"http://localhost:8000/v1\", api_key=\"EMPTY\")\n\nresp = client.chat.completions.create(\n model=\"JetBrains/Mellum2-12B-A2.5B-Instruct\",\n messages=[{\"role\": \"user\", \"content\": \"Write a Python function to reverse a string.\"}],\n max_tokens=81920,\n temperature=0.6,\n top_p=0.95,\n extra_body={\"top_k\": 20},\n)\nprint(resp.choices[0].message.content)\n```\n\n## References\n\n- [Model card](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Instruct)\n- [vLLM support PR #43992](https://github.com/vllm-project/vllm/pull/43992)\n- [Mellum2 Technical Report](https://arxiv.org/abs/2605.31268)\n", "source": { "source": "vllm", "repo": "vllm-project/vllm", "ref": "main", "sha": "389ffebe555705c85337025d5df108d0384d365e", "path": "recipes/models/JetBrains/Mellum2-12B-A2.5B-Instruct.yaml", "edit_url": "https://github.com/vllm-project/vllm/edit/main/recipes/models/JetBrains/Mellum2-12B-A2.5B-Instruct.yaml", "overlays": [] } }