xiaomi Open weights Reasoning

MiMo-V2.5-Pro pricing & benchmarks

Released Apr 22, 2026 · served by 6 providers · ranked #23 by intelligence in our index

Input / 1M tokens$0.348
Output / 1M tokens$0.696
Context window1.1M
Cached input$0.0036
Intelligence index42.2
Coding index60.2
Agentic index29.1
1M in + 300K out$0.56

What MiMo-V2.5-Pro is

MiMo-V2.5-Pro is Xiaomi’s flagship model, delivering strong performance in general agentic capabilities, complex software engineering, and long-horizon tasks, with top rankings on benchmarks such as ClawEval, GDPVal, and SWE-bench Pro....

Providers and prices for MiMo-V2.5-Pro

The same model costs different amounts depending on who serves it. Prices are per 1M tokens, cheapest first. Uptime and throughput are OpenRouter's rolling measurements, not vendor claims.

ProviderInputOutputContext QuantUptime 24hThroughput
GMICloudcheapest $0.348 $0.696 1.1M bf16 93.1%
Xiaomi $0.435 $0.870 1.0M fp8 99.7%
AtlasCloud $0.435 $0.870 1.0M fp8 97.0%
Novita $0.480 $0.960 1.0M 99.6%
DigitalOcean $0.600 $3.00 262K 99.2%
DeepInfra $1.00 $3.00 1.0M fp8 92.4%

MiMo-V2.5-Pro benchmark results

Design Arena head-to-head results, as reported through the OpenRouter model API.

CategoryArenaRankEloWin rate
gamedev models #5 1339 62%
codecategories models #9 1312 58.1%
website models #11 1309 57.4%
dataviz models #13 1295 58.1%
3d models #15 1309 57.4%
uicomponent models #18 1297 56.3%
svg models #21 1222 51.9%
asciiart models #33 1180 47.1%

Cheaper models in the same class

Models scoring within 4 points of MiMo-V2.5-Pro on the Intelligence Index, but with a lower output price.

Compare MiMo-V2.5-Pro head-to-head

FAQ

How much does the MiMo-V2.5-Pro API cost?

$0.348 per 1M input tokens and $0.696 per 1M output tokens. A typical workload of 1M input + 300K output tokens costs about $0.56.

Is MiMo-V2.5-Pro free?

No. It is a paid model starting at $0.348 per 1M input tokens, though some providers offer trial credits.

Which provider is cheapest for MiMo-V2.5-Pro?

GMICloud at $0.348 input / $0.696 output per 1M tokens (bf16 quantization).

Can I self-host MiMo-V2.5-Pro?

Yes — weights are published as XiaomiMiMo/MiMo-V2.5-Pro on Hugging Face, so you can run it on your own hardware.