‹ PublicAI Index
The LLM benchmark aggregator.
GPT-6.1 Sol
OpenAI
Strongest in Language (#2 of 62), weakest in Legal (#100 of 164). Above par in 25 of 27 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5.1 and Claude Fable 5 and ahead of Gemini 4 Argon and Claude Opus 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.
Core abilities64.6−0.6#3/2253/4
Language65.8−0.8#2/621/1
Data analysis62.9−3.2#3/481/1
General intelligence66.5−2.3#7/2142/3
Instruction following61−12.8#12/621/1
Reasoning64.2−0.7#4/1905/5
Reasoning65.8−0.9#4/1443/3
Mathematics63.7−0.4#5/1502/2
Agents63.3−2.6#7/2893/6
Long-horizon work69.4−4.6#5/481/1
Web research58.9−3.7#9/491/2
Computer use60.2−7.5#12/811/2
Knowledge work66.1−5.7#15/1982/2
Coding61.6−4.8#8/1764/6
Agentic coding62.1−4.9#8/1694/5
Knowledge57.3−5.4#22/1911/3
Document parsing57.8−5.8#21/771/1
Human preference64.3−4.3#25/3501/1
Human preference64.3−4.3#25/3501/1
Professional54.4−8.7#37/1882/2
IT operations73.8leads#2/311/1
Biology research66−2.2#2/281/1
Document design60.9−6.4#10/481/1
Medical57.5−9#24/1491/1
Cybersecurity26.6−37.4#45/461/1
Finance52.9−9.6#75/1611/1
Legal48−17.9#100/1641/1
§ 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 GPT-6.1 Sol, left for the other.
§ 3 · Sources
Where the numbers come from
11 publications, 40 figures. Every one links to the page it was read from.
LMArena Text 1479
Artificial Analysis Intelligence Index 52
GDPval-AA 1592
AA-Briefcase 1556
Terminal-Bench 58.2% (Codex)
ARC-AGI-2 94.2%
LiveBench 85.4LiveBench · Agentic 73LiveBench · Reasoning 92.6LiveBench · Coding 96LiveBench · Browser Use 88LiveBench · Data Analysis 82.7LiveBench · Research 83.5LiveBench · Design 82.1
LiveBench 1.0 81.6LiveBench 1.0 · Reasoning 92.6LiveBench 1.0 · Coding 80.4LiveBench 1.0 · Agentic Coding 54.5LiveBench 1.0 · Mathematics 96.8LiveBench 1.0 · Data Analysis 82.7LiveBench 1.0 · Language 90.1LiveBench 1.0 · Instruction Following 74.2
SimpleBench 82.9%
Vals · Legal Research Bench 38.46%Vals · Harvey Legal Agent Benchmark 6.25%Vals · Finance Agent 52.03%Vals · MedCode 48.84%Vals · MedScribe 86.45%Vals · BioMysteryBench 79.63%Vals · CyberBench 39.29%Vals · SRE Bench 50.76%Vals · Vibe Code Bench 88.93% (OpenHands)Vals · Code Migration 65.12%Vals · ProofBench 99%
ParseBench · Overall 70.48%ParseBench · Tables 92.44%ParseBench · Charts 37.16%ParseBench · Content faithfulness 89.25%ParseBench · Semantic formatting 74.78%ParseBench · Visual grounding 58.76%
Badge
[](https://publicai.io/model-index/m/gpt-6-1-sol)