Intel / post

Heads up, agent users！ If you're using Xiaomi MiMo with thinking mode: When…

Source: x.com
Date: Published May 12, 2026

Heads up, agent users！ If you're using Xiaomi MiMo with thinking mode: When thinking mode is enabled in a multi-turn agent session and the conversation history contains a tool call, any assistant message with tool calls passed back in subsequent user turns must preserve its full reasoning_content field — otherwise the API will return a 400 error. Without it, the model's context is incomplete, which can lead to weaker instruction-following, more hallucinations, and a visibly degraded user experience. Missing reasoning = incomplete context = degraded reasoning quality. Affected frameworks include TRAE, Cursor, Roo Code, Codex, GitHub Copilot CLI, Zed, AutoGen. We're actively working with the maintainers to push compatibility updates. Affected models: MiMo-V2.5-Pro, MiMo-V2.5, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2-Flash. See docs(https://t.co/9v7oybaW26 )for more details.

May 12, 2026 View on X

Context

Xiaomi MiMo posted a compatibility warning: when thinking mode is enabled in a multi-turn session, Xiaomi says any assistant message containing a tool call must carry its full reasoning_content field forward when that turn is passed back into a later user turn, or the API returns a 400 error. Xiaomi lists the affected models as MiMo-V2.5-Pro, MiMo-V2.5, MiMo-V2-Pro, MiMo-V2-Omni and MiMo-V2-Flash, and names TRAE, Cursor, Roo Code, Codex, GitHub Copilot CLI, Zed and AutoGen as affected frameworks, meaning the gap sits in how those tools reconstruct conversation history rather than in the model itself. Xiaomi frames this as an active, unresolved compatibility gap it is working through with those maintainers, not something already fixed, so anyone using thinking mode through one of those frameworks may still hit the described errors or degraded output.

Terms in this piece · Glossary

AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.

Read the source x.com