- input
- $2.50
- cached
- $1.25
- output
- $10.00
- Context
- 128Ktokens
- Released
- 1y ago
Cost/1M tokens
Benchmarks
DeepSWE is the primary coding benchmark where an exact model evaluation exists. Arena leaderboards and other sourced evaluations appear below; missing scores are not zero.
How it ranks
Against its predecessor, similarly priced models and the models ranked around it.
DeepSWE 1.1
Best pass@1 per model · Higher is better
Not evaluated in DeepSWE 1.1
Arena Text
Best Arena rating per model · Higher is better
Not ranked on Arena Text yet
Arena WebDev
Best Arena rating per model · Higher is better
Not ranked on Arena WebDev yet
Arena Vision
Best Arena rating per model · Higher is better
Not ranked on Arena Vision yet
Arena Search
Best Arena rating per model · Higher is better
Not ranked on Arena Search yet
Arena Document
Best Arena rating per model · Higher is better
Not ranked on Arena Document yet
Agent Arena
Best IPS score per model · Higher is better
Not ranked on Agent Arena yet
ARC-AGI-2
Best score per model · Higher is better
#78 of 85
- 83.9%
Claude Sonnet 5.5Similar price - 77.1%
Gemini 3.1 Pro PreviewSimilar price - 67.1%
Grok 4.6Similar price - 60.4%
Kimi K3Similar price - 0.0%
GPT-4.1 MiniNearby rank - 0.0%GPT-4.1 NanoNearby rank
- 0.0%GPT-4o (2024-11-20)
- 0.0%GPT-4o-mini (2024-07-18)Nearby rank
Ahead of 0 of 4 similarly priced models
FrontierMath-2025-02-28-Private
Best score per model · Higher is better
#54 of 58
- 36.9%Gemini 3.1 Pro PreviewSimilar price
- 32.4%Claude Sonnet 4.6Similar price
- 14.1%Gemini 2.5 ProSimilar price
- 10.3%Gemini 2.5 Pro Preview 06-05Similar price
- 0.3%
Mistral Large 2.1Nearby rank - 0.3%GPT-4o (2024-08-06)Nearby rank
- 0.3%GPT-4o (2024-11-20)
- 0.0%
Llama 4 ScoutNearby rank
Ahead of 0 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#140 of 161
- 95.6%Claude Sonnet 5.5Similar price
- 94.4%Gemini 3.1 Pro PreviewSimilar price
- 94.0%Grok 4.6Similar price
- 93.1%Kimi K3Similar price
- 48.9%GPT-4o (2024-05-13)Nearby rank
- 48.1%
Qwen-PlusNearby rank - 47.9%GPT-4o (2024-11-20)
- 47.7%Gemma 3 27BNearby rank
- 47.6%Magistral Small 1.2Nearby rank
Ahead of 0 of 4 similarly priced models
Humanity's Last Exam
Best score per model · Higher is better
#34 of 34
- 46.4%Gemini 3.1 Pro PreviewSimilar price
- 46.1%Claude Sonnet 5.5Similar price
- 21.6%Gemini 2.5 Pro Preview 06-05Similar price
- 7.5%Claude Sonnet 4.5Similar price
- 4.5%Mistral Medium 3Nearby rank
- 4.1%Claude Sonnet 3.5 v2Nearby rank
- 2.7%GPT-4o (2024-11-20)
Ahead of 0 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#144 of 158
- 100.0%Claude Sonnet 5.5Similar price
- 100.0%Qwen3.8 Max (0902)Similar price
- 97.2%Kimi K3Similar price
- 95.6%Gemini 3.1 Pro PreviewSimilar price
- 6.4%GPT-4o (2024-08-06)Nearby rank
- 6.3%GPT-4o (2024-05-13)Nearby rank
- 6.3%GPT-4o (2024-11-20)
- 6.1%
Qwen TurboNearby rank - 5.1%Llama 3.3 70B InstructNearby rank
Ahead of 0 of 4 similarly priced models
SWE-bench Verified
Best score per model · Higher is better
#35 of 35
- 77.3%Qwen3.7 MaxSimilar price
- 75.6%Gemini 3.1 Pro Preview Custom ToolsSimilar price
- 75.2%Claude Sonnet 4.6Similar price
- 57.6%Gemini 2.5 ProSimilar price
- 48.5%GPT-4.1Nearby rank
- 31.0%GPT-4o (2024-11-20)
Ahead of 0 of 4 similarly priced models
Not evaluated in DeepSWE 1.1 as of 2026-09-22. Results are only shown for exact model versions in the official dataset. Official leaderboard.
All published scores
- ARC-AGI-2#78 of 850.0%
- default0.0%
- FrontierMath-2025-02-28-Private#54 of 580.3%
- default0.3%
Source:Epoch AI evaluation - GPQA diamond#140 of 16147.9%
- default47.9%
Source:Epoch AI evaluation - Humanity's Last Exam#34 of 342.7%
- default2.7%
- OTIS Mock AIME 2024-2025#144 of 1586.3%
- default6.3%
Source:Epoch AI evaluation - SWE-bench Verified#35 of 3531.0%
- default31.0%
Source:Epoch AI evaluation
Benchmark compilation: Epoch AI, Capabilities & benchmarking (CC BY attribution). Scores are converted to percentages and model names matched to the Vertical Tabs catalog. Individual evaluations are credited to their original sources. External results may carry separate licences and evaluation configurations; do not interpret scores from different harnesses as directly comparable.