Open Source

Openrouter stealth model Hunter/Healer Alpha has been officially confirmed as MiMo, and a new model is coming.

The viral 'Hunter Alpha' is MiMo V2 Pro, a text-only reasoning model with a massive 1M token context.

Deep Dive

The AI community's speculation is over. The mysterious 'Hunter Alpha' and 'Healer Alpha' models that recently appeared on Openrouter have been officially identified through a GitHub pull request. Hunter Alpha is the MiMo V2 Pro, a specialized text-only reasoning model boasting an exceptionally large 1,048,576 token context window. Its counterpart, Healer Alpha, is the MiMo V2 Omni, a multimodal model capable of both text and image reasoning with a 262,144 token context. Both models share a maximum generation limit of 32,000 tokens, positioning them as high-capacity tools for complex, long-context tasks.

This confirmation from the OpenClaw repository solidifies MiMo's place as a serious contender in the reasoning model space, moving from speculation to a concrete offering. The separation into a pure-text 'Pro' variant and a multimodal 'Omni' variant suggests a strategic focus on both raw reasoning power and versatile understanding. For developers on Openrouter, this means immediate access to two new, high-spec models capable of handling lengthy documents, detailed analysis, and integrated visual reasoning, all through a single API endpoint. The announcement also teases that a new model is 'coming,' indicating this is just the beginning of MiMo's rollout.

Key Points
  • Hunter Alpha is MiMo V2 Pro: a text-only reasoning model with a 1,048,576 token (1M) context window.
  • Healer Alpha is MiMo V2 Omni: a text + image reasoning model with a 262,144 token context window.
  • Both models have a max output of 32,000 tokens and are available now on the Openrouter platform.

Why It Matters

Developers gain immediate API access to powerful, long-context reasoning models for complex analysis and multimodal tasks.