‹ PublicAI Index
The LLM benchmark aggregator.
Kimi K3.1 Pro
Moonshot · 2.8T · open weights
Strongest in Knowledge work (#5 of 208), weakest in Legal (#81 of 173). Above par in 20 of 20 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 5.5 and Gemini 4 Carbon and ahead of Gemini 3.1 Pro and Grok 4.6.
§ 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.
Agents58.7−7.6#21/2983/6
Knowledge work69.1−1.6#5/2082/2
Long-horizon work50.1−23.7#15/321/1
Web research52.5−11.1#17/361/2
Computer use56.3−11.8#23/691/2
Human preference64.6−4.2#23/3541/1
Human preference64.6−4.2#23/3541/1
Reasoning57.5−7.9#25/1952/5
Mathematics58.7−5.1#17/1521/2
Reasoning57.4−11.6#36/1471/3
Core abilities56.7−8.9#32/2002/4
Data analysis56.7−9.4#13/321/1
General intelligence58−11#37/1802/3
Coding53.1−13#55/1792/6
Agentic coding53.3−13.4#54/1732/5
Professional53.2−9.2#56/1942/2
Biology research58.7−12#8/311/1
IT operations59.5−14.3#9/331/1
Document design50.7−17.7#18/321/1
Legal51.4−14.5#81/1731/1
§ 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 Kimi K3.1 Pro, left for the other.
§ 3 · Sources
Where the numbers come from
6 publications, 18 figures. Every one links to the page it was read from.
LMArena Text 1483
Artificial Analysis Intelligence Index 46
GDPval-AA 1861
AA-Briefcase 1674
LiveBench 68.7LiveBench · Agentic 38.7LiveBench · Reasoning 91.1LiveBench · Coding 67.8LiveBench · Browser Use 75.9LiveBench · Data Analysis 79.1LiveBench · Research 82.1LiveBench · Design 73.7
Vals · Legal Research Bench 37.5%Vals · Harvey Legal Agent Benchmark 11.67%Vals · BioMysteryBench 75.56%Vals · SRE Bench 27.1%Vals · Vibe Code Bench 84.89% (OpenHands)Vals · ProofBench 91%
Badge
[](https://publicai.io/model-index/m/kimi-k3-1-pro)