Xiaomi: mimo-v2.5

mimo-v2.5
1M context131.1K outReasoningToolsVisionAudio inVideoCacheStructured
Released Apr 22, 2026Knowledge cutoff 2026Updated Apr 22, 2026

MiMo-V2.5 is the versatile, high-efficiency backbone of Xiaomi's 2.5-series AI ecosystem. While inheriting the "Harness-Awareness" architecture from the Pro version, MiMo-V2.5 is optimized for lower latency and extreme token efficiency in repetitive, high-frequency autonomous tasks. It excels at managing medium-to-long execution trajectories (up to 100+ tool calls) and maintaining precise context alignment. In comparative testing, it achieved a 50% reduction in token consumption compared to previous generation models like Muse Spark while maintaining superior task success rates. It is the ideal engine for 2026's decentralized AI micro-services and real-time coding assistants.

Mode chatTokenizer OtherQuantization fp8XiaomiMiMo/MiMo-V2.5

Pricing

Input price
$0.20/ 1M tokens
Output price
$0.39/ 1M tokens
Context window 262.1K tokensCompatible endpoints openaiVendor Xiaomi

Performance

Loading performance data...

Usage & Ranking

Loading usage...

Supported parameters

All providers = every upstream serving this model supports it. Some providers = depends on which upstream handles the request. Default = the value sent when you leave the parameter unset.

ParameterProvidersDefault
frequency_penaltySome providersNot sent by default
include_reasoningAll providers-
logit_biasSome providers-
logprobsSome providers-
max_tokensAll providers-
min_pSome providers-
presence_penaltySome providersNot sent by default
reasoningAll providers-
repetition_penaltySome providersNot sent by default
response_formatAll providers-
seedSome providers-
stopSome providers-
structured_outputsSome providers-
temperatureAll providers1
tool_choiceAll providers-
toolsAll providers-
top_kSome providersNot sent by default
top_logprobsSome providers-
top_pAll providers0.95

Frequently asked questions

Similar models