Xiaomi open-sources AI coding assistant MiMo Code with enhanced performance

icon MarsBit
Share
Share IconShare IconShare IconShare IconShare IconShare IconCopy
AI summary iconSummary

expand icon
Xiaomi has open-sourced MiMo Code V0.1.0 under the MIT license, integrating the MiMo-V2.5 model and supporting major large models such as DeepSeek and Kimi. The tool features a persistent memory system and a /dream command for optimization. It achieved 62% on SWE-Bench Pro and 73% on Terminal Bench 2, outperforming Claude Code. Voice control and cross-platform deployment via curl or npm are supported. With growing interest in AI tools, this move reflects increasing developer demand. The fear and greed index in the tech sector remains elevated as open-source projects gain momentum.

According to Beating Monitor, Xiaomi MiMo has officially released and open-sourced the terminal AI programming assistant MiMo Code V0.1.0 under the permissive MIT open-source license. MiMo Code is developed based on the open-source project OpenCode, featuring a time-limited free MiMo-V2.5 multimodal model and compatibility with major large model APIs such as DeepSeek, Kimi, and GLM, as well as third-party Token Plans. Traditional programming agents often rely on models to autonomously take notes, frequently forgetting critical context due to the model’s failure to trigger active recall. MiMo Code introduces a persistent memory system that outsources state logging to independent subagents. When a conversation window approaches its limit, the subagent automatically saves the state and generates a clean summary for the main agent to seamlessly continue. The built-in /dream command runs automatically every seven days, with a dedicated agent reading historical conversations and memory files to perform merging, deduplication, and path validity checks, compressing scattered information into the current state to update global memory. To enhance compatibility with model and agent frameworks, MiMo Code features a dedicated Harness system designed specifically for the MiMo series models. Users can press the Tab key to switch to Compose mode; after providing basic requirements, the system autonomously executes the full development cycle—including design, planning, coding, testing, and review. In authoritative benchmarks SWE-Bench Pro and Terminal Bench 2, MiMo Code achieved scores of 62% and 73% respectively using the same base model, outperforming Claude Code by 5 percentage points. MiMo Code also includes voice control functionality, allowing users to modify inputs or execute actions such as sending via voice commands. For deployment and usage, macOS and Linux users can install with a single curl command, while Windows users can deploy via npm. Upon launch, the terminal displays a fully localized TUI interface with a persistent status dashboard on the right side for real-time monitoring of progress.

Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.