- input
- $30.00
- output
- $180.00
- Context
- 1.1Mtokens
- Released
- 8mo 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
#18 of 85
- 84.6%
GPT-5.5 ProSimilar price - 83.9%GPT-5.6 TerraNearby rank
- 83.9%
Claude Sonnet 5.5Nearby rank - 83.3%GPT-5.4 Pro
- 81.7%
Inkling 2 SmallNearby rank - 78.1%
DeepSeek V4.1 Flash 1017Nearby rank - 54.2%GPT-5.2 ProPredecessor
- 18.3%GPT-5 ProSimilar price
- 4.9%o3 ProSimilar price
+29.2 pts vs GPT-5.2 Pro (predecessor) · Ahead of 2 of 3 similarly priced models
FrontierMath-2025-02-28-Private
Best score per model · Higher is better
#1 of 58
- 50.0%GPT-5.4 Pro
- 47.6%GPT-5.4Nearby rank
- 47.2%Claude Opus 4.8Nearby rank
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#19 of 72
- 78.0%GPT-5.5 ProSimilar price
- 70.7%GPT-5.6 TerraNearby rank
- 61.0%GPT-5.6 LunaNearby rank
- 58.5%GPT-5.4 Pro
- 56.1%Claude Opus 4.8Nearby rank
- 56.1%GPT-6 LunaNearby rank
- 46.0%GPT-5.2 ProPredecessor
- 19.5%GPT-5 ProSimilar price
+12.5 pts vs GPT-5.2 Pro (predecessor) · Ahead of 1 of 2 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#17 of 87
- 87.7%GPT-5.5 ProSimilar price
- 85.6%Claude Opus 5Nearby rank
- 85.3%GPT-5.5Nearby rank
- 82.5%GPT-5.4 Pro
- 82.1%GPT-5.6 LunaNearby rank
- 80.0%Claude Opus 4.8Nearby rank
- 74.0%GPT-5.2 ProPredecessor
- 55.8%GPT-5 ProSimilar price
+8.5 pts vs GPT-5.2 Pro (predecessor) · Ahead of 1 of 2 similarly priced models
GPQA diamond
Best score per model · Higher is better
#8 of 161
- 95.4%GPT-6.1 SolNearby rank
- 94.8%
Gemini 3.7 FlashNearby rank - 94.6%GPT-5.4 Pro
- 94.4%Gemini 3.1 Pro PreviewNearby rank
- 94.3%GPT-6 SolNearby rank
Humanity's Last Exam
Best score per model · Higher is better
#6 of 34
- 46.1%Claude Sonnet 5.5Nearby rank
- 44.5%Gemini 3.8 FlashNearby rank
- 44.3%GPT-5.4 Pro
- 37.5%Gemini 3 Pro PreviewNearby rank
- 36.2%GPT-5.4Nearby rank
- 31.6%GPT-5 ProEarlier version
+12.7 pts vs GPT-5 Pro (earlier version)
SimpleQA Verified
Best score per model · Higher is better
#40 of 87
- 47.0%DeepSeek V4 Pro 0423Nearby rank
- 46.5%Claude Sonnet 5.5Nearby rank
- 46.3%GPT-5.4 Pro
- 45.8%
Qwen3.8 MaxNearby rank - 45.7%Claude Opus 4.5Nearby rank
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#18 of 8583.3%
+29.2 pts vs GPT-5.2 Pro (predecessor)
- xhigh83.3%
- FrontierMath-2025-02-28-Private#1 of 5850.0%
- xhigh50.0%
Source:Epoch AI evaluation - FrontierMath-Tier-4-v2-Private#19 of 7258.5%
+12.5 pts vs GPT-5.2 Pro (predecessor)
- xhigh58.5%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#17 of 8782.5%
+8.5 pts vs GPT-5.2 Pro (predecessor)
- xhigh82.5%
Source:Epoch AI evaluation - GPQA diamond#8 of 16194.6%
- xhigh94.6%
Source:Epoch AI evaluation - Humanity's Last Exam#6 of 3444.3%
+12.7 pts vs GPT-5 Pro (earlier version)
- unknown44.3%
- SimpleQA Verified#40 of 8746.3%
- xhigh46.3%
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.