- Context
- 1Mtokens
- Released
- 9mo 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
Arena #6
- 1511
Muse WonderSimilar price - 1506
Claude Opus 5.5Nearby rank - 1504Claude Fable 5Nearby rank
- 1503Claude Opus 4.6
- 1503Claude Fable 5.1Nearby rank
- 1501Claude Opus 4.7Nearby rank
- 1485
GPT-5.6 SolSimilar price - 1481GPT-5.5Similar price
- 1474Claude Opus 4.5Predecessor
- 1346GPT-4o (2024-05-13)Similar price
+29 vs Claude Opus 4.5 (predecessor) · Ahead of 3 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #52
- 1649Muse WonderSimilar price
- 1615GPT-5.6 SolSimilar price
- 1555Claude Opus 4.8Nearby rank
- 1552
Grok 4.5Nearby rank - 1547Claude Opus 4.6
- 1541Muse Spark 1.1Nearby rank
- 1539Claude Sonnet 5Nearby rank
- 1512GPT-5.5Similar price
- 1493Claude Opus 4.5Predecessor
- 1462GPT-5.4Similar price
+54 vs Claude Opus 4.5 (predecessor) · Ahead of 2 of 4 similarly priced models
Arena Vision
Best Arena rating per model · Higher is better
Arena #4
- 1300Claude Opus 4.7Nearby rank
- 1299
Qwen3.8 MaxNearby rank - 1299Claude Opus 4.6
- 1295Claude Opus 5.5Nearby rank
- 1294
Gemini 3.7 FlashNearby rank - 1288GPT-5.6 SolSimilar price
- 1285GPT-5.5Similar price
- 1282GPT-5.4Similar price
- 1208Claude Opus 4Earlier version
- 1162GPT-4o (2024-05-13)Similar price
+91 vs Claude Opus 4 (earlier version) · Ahead of 3 of 4 similarly priced models
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
Arena #3
- 1500Claude Fable 5Nearby rank
- 1498Claude Fable 5.1Nearby rank
- 1497Claude Opus 4.6
- 1494Claude Opus 4.7Nearby rank
- 1486GPT-5.6 SolSimilar price
- 1481GPT-5.5Similar price
- 1474GPT-6 AstraSimilar price
- 1471GPT-5.4Similar price
- 1467Claude Opus 4.5Predecessor
+29 vs Claude Opus 4.5 (predecessor) · Ahead of 3 of 4 similarly priced models
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
#30 of 85
- 95.0%GPT-6 AstraSimilar price
- 92.5%GPT-5.6 SolSimilar price
- 85.0%GPT-5.5Similar price
- 74.0%GPT-5.4Similar price
- 65.8%
GLM 5.3 FlashNearby rank - 65.1%Grok 4.20Nearby rank
- 64.6%Claude Opus 4.6
- 61.4%
DeepSeek V4 Flash 0731Nearby rank - 61.4%Grok 4.7Nearby rank
- 37.6%Claude Opus 4.5Predecessor
+26.9 pts vs Claude Opus 4.5 (predecessor) · Ahead of 0 of 4 similarly priced models
FrontierMath-2025-02-28-Private
Best score per model · Higher is better
#5 of 58
- 47.6%GPT-5.4Similar price
- 47.2%Claude Opus 4.8Similar price
- 43.8%Claude Opus 4.7Similar price
- 40.7%Claude Opus 4.6
- 40.7%GPT-5.2Nearby rank
- 39.0%Gemini 3.5 FlashNearby rank
- 32.4%Claude Sonnet 4.6Similar price
- 20.7%Claude Opus 4.5Predecessor
+20.0 pts vs Claude Opus 4.5 (predecessor) · Ahead of 1 of 4 similarly priced models
FrontierMath-Tier-4-2025-07-01-Private
Best score per model · Higher is better
#4 of 46
- 31.3%Claude Opus 4.8Similar price
- 27.1%GPT-5.4Similar price
- 22.9%Claude Opus 4.7Similar price
- 22.9%Claude Opus 4.6
- 18.8%GPT-5.2Nearby rank
- 18.8%Gemini 3 Pro PreviewNearby rank
- 8.3%Claude Sonnet 4.6Similar price
- 4.2%Claude Opus 4.5Predecessor
+18.7 pts vs Claude Opus 4.5 (predecessor) · Ahead of 1 of 4 similarly priced models
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#38 of 72
- 97.6%GPT-6 AstraSimilar price
- 82.9%GPT-5.6 SolSimilar price
- 72.5%GPT-5.5Similar price
- 49.0%GPT-5.4Similar price
- 29.3%GLM 5.2Nearby rank
- 29.3%GLM 5.3Nearby rank
- 26.8%Claude Opus 4.6
- 26.8%DeepSeek V4 Pro 0813Nearby rank
- 26.8%DeepSeek V4.1 FlashNearby rank
- 4.9%Claude Opus 4.5Predecessor
+22.0 pts vs Claude Opus 4.5 (predecessor) · Ahead of 0 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#34 of 87
- 93.7%GPT-6 AstraSimilar price
- 89.1%GPT-5.6 SolSimilar price
- 85.3%GPT-5.5Similar price
- 78.6%GPT-5.4Similar price
- 67.4%GPT-5.2Nearby rank
- 67.4%DeepSeek V4.1 FlashNearby rank
- 66.0%Claude Opus 4.6
- 66.0%Grok 4.6Nearby rank
- 65.6%Claude Sonnet 5Nearby rank
- 34.4%Claude Opus 4.5Predecessor
+31.6 pts vs Claude Opus 4.5 (predecessor) · Ahead of 0 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#40 of 161
- 93.5%GPT-5.6 SolSimilar price
- 93.3%GPT-5.4Similar price
- 90.7%GPT-5.5Similar price
- 90.6%Claude Opus 5.5Nearby rank
- 90.5%Claude Opus 4.6
- 90.5%Claude Sonnet 5Nearby rank
- 90.5%GPT-6 LunaNearby rank
- 86.0%Claude Opus 4.5Predecessor
- 48.9%GPT-4o (2024-05-13)Similar price
+4.5 pts vs Claude Opus 4.5 (predecessor) · Ahead of 1 of 4 similarly priced models
Humanity's Last Exam
Best score per model · Higher is better
#10 of 34
- 54.8%GPT-6 AstraSimilar price
- 36.2%GPT-5.4Similar price
- 36.2%Claude Opus 4.7Similar price
- 34.4%Claude Opus 4.6
- 31.6%GPT-5 ProNearby rank
- 30.1%Claude Haiku 5.5Nearby rank
- 14.2%Claude Opus 4.5Predecessor
- 7.5%Claude Sonnet 4.5Similar price
+20.3 pts vs Claude Opus 4.5 (predecessor) · Ahead of 1 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#43 of 158
- 100.0%GPT-5.6 SolSimilar price
- 97.8%GPT-5.4Similar price
- 95.6%
Qwen3.7 MaxNearby rank - 94.7%Claude Sonnet 5Nearby rank
- 94.4%Claude Opus 4.6
- 94.4%DeepSeek V4 Flash 0731Nearby rank
- 94.2%Gemini 3.6 FlashNearby rank
- 86.1%Claude Opus 4.5Predecessor
- 84.4%GPT-5.5Similar price
- 6.3%GPT-4o (2024-05-13)Similar price
+8.3 pts vs Claude Opus 4.5 (predecessor) · Ahead of 2 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#37 of 87
- 75.6%GPT-6 AstraSimilar price
- 69.7%GPT-5.6 SolSimilar price
- 63.0%GPT-5.5Similar price
- 48.0%GPT-5.1Nearby rank
- 47.3%Qwen3.8 Max (0902)Nearby rank
- 47.0%Claude Opus 4.6
- 47.0%DeepSeek V4 Pro 0423Nearby rank
- 46.5%Claude Sonnet 5.5Nearby rank
- 45.7%Claude Opus 4.5Predecessor
- 45.1%GPT-5.4Similar price
+1.3 pts vs Claude Opus 4.5 (predecessor) · Ahead of 1 of 4 similarly priced models
SWE-bench Verified
Best score per model · Higher is better
#12 of 35
- 83.5%Claude Opus 4.7Similar price
- 76.9%GPT-5.4Similar price
- 76.7%Claude Opus 4.5Predecessor
- 76.7%
Kimi K2.6Nearby rank - 75.6%Claude Opus 4.6
- 75.6%Gemini 3.1 Pro Preview Custom ToolsNearby rank
- 75.4%Gemini 3 Flash PreviewNearby rank
- 75.2%Claude Sonnet 4.6Similar price
- 71.3%Claude Sonnet 4.5Similar price
-1.0 pts vs Claude Opus 4.5 (predecessor) · Ahead of 2 of 4 similarly priced models
Terminal Bench
Best score per model · Higher is better
#5 of 49
- 84.7%GPT-5.5Similar price
- 81.8%GPT-5.4Similar price
- 80.2%Claude Opus 4.7Similar price
- 80.2%Gemini 3.1 Pro PreviewNearby rank
- 79.8%Claude Opus 4.6
- 78.4%GPT-5.3-CodexNearby rank
- 69.4%Gemini 3 Pro PreviewNearby rank
- 63.1%Claude Opus 4.5Predecessor
- 53.4%Claude Sonnet 4.6Similar price
+16.7 pts vs Claude Opus 4.5 (predecessor) · Ahead of 1 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
- Arena TextArena #61503
+29 vs Claude Opus 4.5 (predecessor)
- high150395% CI 1500 to 1506 · 84,424 votes
- default149695% CI 1493 to 1500 · 89,228 votes
- Arena WebDevArena #521547
+54 vs Claude Opus 4.5 (predecessor)
- high154795% CI 1541 to 1552 · 19,059 votes
- default153795% CI 1531 to 1542 · 20,368 votes
- Arena VisionArena #41299
+91 vs Claude Opus 4 (earlier version)
- high129995% CI 1292 to 1305 · 22,116 votes
- default129395% CI 1287 to 1299 · 26,627 votes
- Arena DocumentArena #31497
+29 vs Claude Opus 4.5 (predecessor)
- high149795% CI 1490 to 1503 · 29,456 votes
- default149695% CI 1490 to 1501 · 47,264 votes
- ARC-AGI-2#30 of 8564.6%
+26.9 pts vs Claude Opus 4.5 (predecessor)
- default64.6%120K
- FrontierMath-2025-02-28-Private#5 of 5840.7%
+20.0 pts vs Claude Opus 4.5 (predecessor)
- max40.7%
- default40.0%32K
- default39.7%64K
- default38.3%
Source:Epoch AI evaluation - FrontierMath-Tier-4-2025-07-01-Private#4 of 4622.9%
+18.7 pts vs Claude Opus 4.5 (predecessor)
- max22.9%
- default20.8%32K
- default20.8%64K
- default14.6%
Source:Epoch AI evaluation - FrontierMath-Tier-4-v2-Private#38 of 7226.8%
+22.0 pts vs Claude Opus 4.5 (predecessor)
- max26.8%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#34 of 8766.0%
+31.6 pts vs Claude Opus 4.5 (predecessor)
- max66.0%
Source:Epoch AI evaluation - GPQA diamond#40 of 16190.5%
+4.5 pts vs Claude Opus 4.5 (predecessor)
- max88.4%
- default90.5%32K
- default88.8%64K
Source:Epoch AI evaluation - Humanity's Last Exam#10 of 3434.4%
+20.3 pts vs Claude Opus 4.5 (predecessor)
- max34.4%
- default19.0%
- OTIS Mock AIME 2024-2025#43 of 15894.4%
+8.3 pts vs Claude Opus 4.5 (predecessor)
- max91.1%
- default94.4%64K
- default93.1%32K
Source:Epoch AI evaluation - SimpleQA Verified#37 of 8747.0%
+1.3 pts vs Claude Opus 4.5 (predecessor)
- max47.0%
Source:Epoch AI evaluation - SWE-bench Verified#12 of 3575.6%
-1.0 pts vs Claude Opus 4.5 (predecessor)
- default75.6%
Source:Epoch AI evaluation - Terminal Bench#5 of 4979.8%
+16.7 pts vs Claude Opus 4.5 (predecessor)
- unknown79.8%ForgeCodehttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown76.4%Meta-Harnesshttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown75.3%Capyhttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown74.7%Terminus-KIRAhttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown72.1%MAYA-V2https://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown71.9%TongAgentshttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- default69.9%DroidTerminal-Bench v2 LeaderboardOriginal-source terms apply
- unknown69.9%Droidhttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown66.9%Cruxhttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown66.5%Muxhttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown62.9%Terminus 2https://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
- unknown58.0%Claude Codehttps://www.tbench.ai/leaderboard/terminal-bench/2.0Original-source terms apply
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.
Arena ratings: Arena (lmarena-ai/leaderboard-dataset), licensed CC BY 4.0. Ratings come from pairwise human votes and are only comparable within one leaderboard; the interval is Arena's 95% confidence interval. Model names are matched to the Vertical Tabs catalog.