MiMo v2.6 Pro
MiMo v2.6 ProAlternatives?
CheaperGPT-6 Lunaxhigh+27%-68%$0.042Function calling only with none · Chat Completions+3
OpenDeepSeek V4 Pro 0813max+41%+406%$0.67Memory (4-bit): 615 GB → 990 GB+6CheaperGPT-5.6 Lunaxhigh+22%-36%$0.085
CheaperGemini 3.1 Flash Litehigh+4%-55%$0.06
CheaperClaude Haiku 5.5xhighsame-7%$0.12Remove sampling parameters+5GPT-6 Lunaxhigh+27%-68%$0.042Function calling only with none · Chat Completions+3GPT-5.6 Lunaxhigh+22%-36%$0.085Gemini 3.1 Flash Litehigh+4%-55%$0.06Claude Haiku 5.5xhighsame-7%$0.12Remove sampling parameters+5TInkling SmallthinkingOpen-5%-33%$0.089Context: 1M → 256K+1GPT-6 Lunalow+17%-97%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow+14%-93%$0.01GPT-6.1 Sollow+69%same$0.13Remove unsupported parameters+5GPT-6 Sollow+47%same$0.13GPT-5.6 Terralow+25%+8%$0.14Prompts with >272K input tokensDeepSeek V4 Pro 0813maxOpen+41%+406%$0.67Memory (4-bit): 615 GB → 990 GB+6DeepSeek V4.1 Flash 1017maxOpen+23%+97%$0.26Pass back reasoning_content · Chat Completions+4TInklingxhighOpen+19%+217%$0.95 / $4.05Context: 1M → 256K+1
Step 5 PreviewthinkingOpen+19%+670%$1.03web_search not available · Chat Completions+2
Kimi K2.7 CodethinkingOpen+14%+306%$0.54Context: 1M → 256K+2MiMo v2.6 FlashthinkingOpen-23%-53%$0.062Pass back reasoning_content+3MiMo v2 Prothinking-24%—$0.43 / $0.87No image inputMiMo v2.5 ProthinkingOpen-36%-59%$0.054No image inputMiMo v2 Omnithinking-45%—$0.14 / $0.28Context: 1M → 262KMiMo v2 Omni 0327-49%——Context: 1M → 256KMiMo v2.5thinkingOpen-52%-86%$0.019MiMo 7B BaseOpenn/a——Context: 1M → 33K+1MiMo 7B RLOpenn/a——Context: 1M → 33K+1All Xiaomi models →Run it yourself
estimateHistory
3 versionsBenchmarks
in Chat & KnowledgeWatch out 4
Pass back reasoning_content›
In deep-thinking multi-turn tool conversations, return the full reasoning_content in later rounds with tool calls or receive HTTP 400. Source ↗
Forced recommended defaults›
Deep thinking overrides supplied temperature and top_p with 1.0 and 0.95. Source ↗
Avoid answer truncation›
max_completion_tokens includes thinking and the final answer; reserve enough tokens to avoid truncating the answer. Source ↗
MOPD checkpoints address tool-call repetition›
Use the MOPD checkpoints for the tool-call repetition fix; the API has served the updated models since September 25 at 06:00 UTC+8 under unchanged names. Source ↗
FAQ
How much does MiMo v2.6 Pro cost?›
$0.43 per 1M input tokens and $0.87 per 1M output tokens; cached input $0.004 per 1M; cheapest via DeepInfra at $0.43 / $0.87. A typical benchmark task costs about $0.13 at default effort.
Is MiMo v2.6 Pro free to use?›
There is no free API; the cheapest paid price is $0.43 per 1M input tokens. The weights are open, so you can run it yourself.
How good is MiMo v2.6 Pro, and what is it best at?›
#12 of 334 in Chat & Knowledge, #16 of 132 in Coding, #22 of 133 in Agents & Tool Use, #31 of 195 in Writing & Language, #26 of 133 in Vision & Documents. Strongest at: Coding, Agents & work tasks, Reasoning, Long context, Knowledge.
Is MiMo v2.6 Pro better than MiMo v2.5?›
On AA Omniscience · Accuracy, MiMo v2.6 Pro scores 34.9 vs 16.8 — higher. It is 599% more expensive per task.
How can I use MiMo v2.6 Pro?›
Through the API of DeepInfra, GMICloud, Novita, Xiaomi.
What is the context window of MiMo v2.6 Pro?›
1M tokens, with up to 131K output tokens.
When was MiMo v2.6 Pro released? Is there a newer version?›
Released 2026-09-22. It is the current version in its line.
Is MiMo v2.6 Pro open source?›
Yes, the weights are open (MIT licence).
Can I run MiMo v2.6 Pro locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can MiMo v2.6 Pro read images?›
It accepts text and images, audio, video as input.
What is MiMo v2.6 Pro?›
MiMo v2.6 Pro is an open-weights language model from Xiaomi, released Sep 2026, the successor to MiMo v2.5.
What are the best alternatives to MiMo v2.6 Pro?›
GPT-6 Luna (cheaper), DeepSeek V4 Pro 0813 (open weights).