Add JetBrains Mellum2 recipes (Thinking + Instruct)

[esmeetu]

Adds vLLM recipes for JetBrains' Mellum2 family — the reasoning-augmented Thinking checkpoint and its direct-answer Instruct sibling (both 12B total / 2.5B active, 64 experts / 8 active, 131K context, bf16).

Shared details

• Architecture: MoE (MellumForCausalLM), bf16, ~29 GB — fits on a single H200/H100/A100. single_node_tp defaults to TP=1.
• vLLM version: nightly. MellumForCausalLM support merged in vllm-project/vllm#43992 on 2026-06-01, after the latest stable v0.22.0 (2026-05-29), so it is not yet in a tagged release. Both recipes set nightly_required: true.
• Tool calling: --enable-auto-tool-choice --tool-call-parser hermes (both).

Thinking vs Instruct

• Thinking emits <think>...</think> chains before the answer → adds --reasoning-parser qwen3. Suited to complex debugging, planning, agentic/math-heavy tasks.
• Instruct answers directly (no externalized CoT) → no reasoning parser. Lower-latency coding and tool use.

The two recipes cross-link via related_recipes.

🤖 Generated with Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
vllm-recipes	Ready	Preview, Comment	Jun 2, 2026 1:03am

[gemini-code-assist]

Code Review

This pull request adds a configuration file for the new Mellum2-12B-A2.5B-Thinking model by JetBrains and registers JetBrains as a provider. Feedback on the configuration suggests lowering the default --max-model-len from 131072 to 32768 to avoid potential Out-Of-Memory (OOM) errors on standard GPUs. Additionally, it is recommended to correct a likely typo in the Python client example, changing max_tokens from 81920 to 8192.

[gemini-code-assist]

Setting the default --max-model-len to the absolute maximum of 131072 in base_args will cause vLLM to attempt to allocate a massive KV cache on startup. For a model of this size (12B total parameters, ~24 GB in bf16), the KV cache for 131k tokens will require an additional ~8 GB+ of VRAM. On GPUs near the minimum VRAM requirement of 29 GB (or even 32GB/40GB GPUs), this will likely result in an Out-Of-Memory (OOM) error during initialization.\n\nConsider setting a more conservative default (e.g., 32768 or 16384) in base_args to ensure the recipe runs out-of-the-box on standard GPUs, and document in the guide that users can scale it up to 131072 if they have higher-end hardware (like an A100 80GB or H100).

  base_args:
    - "--max-model-len"
    - "32768"

[gemini-code-assist]

The max_tokens parameter in the Python client usage example is set to 81920. This is extremely high for a single chat completion response and is likely a typo for 8192 (8k), which is the typical maximum generation length for reasoning models. Setting it excessively high can lead to client-side validation issues or unexpected behavior if the model gets stuck in a loop.

      max_tokens=8192,