‹ PublicAI Index
The LLM benchmark aggregator.
Mimo V2.6 Pro
Xiaomi · 1T · open weights
Strongest in Cybersecurity (#3 of 9), weakest in Fairness (#172 of 300). Above par in 19 of 20 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5.1 and Muse Spark 1.3 and ahead of Claude Sonnet 5 and GPT-6 Sol.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Agents64−4.1#9/2682/5
Knowledge work68.2−5.4#8/1782/2
Core abilities58.6−8.8#14/2041/3
General intelligence62.4−7.7#11/2041/3
Professional57.3−4.3#15/1681/1
Cybersecurity57.5−2.2#3/91/1
Legal60.5−6.6#10/1511/1
Finance57.2−5.7#26/1511/1
Coding58.2−10.9#15/1651/5
Agentic coding59.5−9.2#14/1571/4
Human preference64.7−2.8#21/3421/1
Human preference64.7−2.8#21/3421/1
Reasoning53.1−13.8#63/1781/4
Mathematics54.7−11#55/1401/2
Safety52.1−9.4#130/3371/3
Jailbreak resistance57.8−7.1#70/2721/1
Toxicity avoidance55.3−3.7#96/2721/1
Secure code53.6−12.9#123/2741/1
Harm refusal51.2−10.4#139/3001/2
Fairness46−24#172/3001/2
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Mimo V2.6 Pro, left for the other.
§ 3 · Sources
Where the numbers come from
6 publications, 17 figures. Every one links to the page it was read from.
LMArena Text 1480
Artificial Analysis Intelligence Index 46
GDPval-AA 1673
AA-Briefcase 1517
Vals · Legal Research Bench 47.12%Vals · Harvey Legal Agent Benchmark 10.83%Vals · Finance Agent 57.34%Vals · CyberBench 72.86%Vals · Vibe Code Bench 85.22% (OpenHands)Vals · Code Migration 43.01%Vals · ProofBench 70%
Enkrypt · Jailbreak risk 6.2%Enkrypt · Harmful content risk 2.2%Enkrypt · CBRN risk 28.3%Enkrypt · Toxicity risk 2.6%Enkrypt · Bias risk 84%Enkrypt · Insecure code risk 26.2%
Badge
[](https://publicai.io/model-index/m/mimo-v2-6-pro)