- input
- $0.70
- cached
- $0.07
- output
- $2.10
- Context
- 1Mtokens
- Released
- 1mo ago
- Size
- 1.1Tparameters
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 #129
- 1458
DeepSeek V4 Pro 0423Similar price - 1450
Kimi K2.5Similar price - 1441
GLM 4.7Similar price - 1426
Nemotron 3 UltraSimilar price - 1424GLM 4.6Nearby rank
- 1423
Qwen3 MaxNearby rank - 1423
Mistral Large 4 - 1422
Claude Haiku 5.5Nearby rank - 1422
Qwen3 235B A22B Instruct 2507Nearby rank - 1415Mistral Large 3 2512Predecessor
Level with Mistral Large 3 2512 (predecessor, overlapping 95% CIs) · Ahead of 0 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #47
- 1632
Hy4 previewSimilar price - 1577
GPT-6 LunaNearby rank - 1573
Step 5 PreviewNearby rank - 1564Mistral Large 4
- 1557Claude Opus 4.7Nearby rank
- 1555Claude Opus 4.8Nearby rank
- 1448DeepSeek V4 Pro 0423Similar price
- 1436Kimi K2.5Similar price
- 1435GLM 4.7Similar price
- 1233Mistral Large 3 2512Predecessor
+331 vs Mistral Large 3 2512 (predecessor) · Ahead of 3 of 4 similarly priced models
Arena Vision
Best Arena rating per model · Higher is better
Arena #73
- 1266
Gemini 3.5 Flash LiteSimilar price - 1252Kimi K2.5Similar price
- 1232
MiniMax M3Nearby rank - 1232GLM 5V TurboNearby rank
- 1231Mistral Large 4
- 1226Qwen3.5-122B-A10BNearby rank
- 1216
MiMo-V2-OmniNearby rank - 1200Mistral Large 3 2512Predecessor
- 1153GLM 4.5VSimilar price
- 1121Qwen2.5 VL 72B InstructSimilar price
+31 vs Mistral Large 3 2512 (predecessor) · Ahead of 2 of 4 similarly priced models
Arena Search
Best Arena rating per model · Higher is better
Not ranked on Arena Search yet
Agent Arena
Best IPS score per model · Higher is better
Arena #44
- +0.005Hy4 previewSimilar price
- -0.022Gemini 3.7 FlashSimilar price
- -0.037Step 5 PreviewSimilar price
- -0.046Qwen3.8 Flash NextNearby rank
- -0.048Qwen3.8 27BNearby rank
- -0.054Mistral Large 4
- -0.059
Beam 1.0 501B A23BNearby rank - -0.072Qwen3.7 MaxNearby rank
- -0.156Nemotron 3 UltraSimilar price
Ahead of 1 of 4 similarly priced models
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#38 of 72
- 26.8%Gemini 3.1 Pro PreviewNearby rank
- 26.8%Gemini 3.5 FlashNearby rank
- 26.8%Mistral Large 4
- 25.6%Kimi K2.6Nearby rank
- 24.4%DeepSeek V4 Flash 0731Nearby rank
- 17.1%Gemini 3 Flash PreviewSimilar price
- 14.6%
Grok 4.3Similar price - 12.2%Kimi K2.7 CodeSimilar price
- 2.4%DeepSeek V4 Pro 0423Similar price
Ahead of 4 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#44 of 87
- 59.2%GLM 5.2Nearby rank
- 58.9%Gemini 3.6 FlashNearby rank
- 58.6%Mistral Large 4
- 57.5%DeepSeek V4 Flash 0731Nearby rank
- 57.2%Kimi K2.6Nearby rank
- 54.0%Kimi K2.7 CodeSimilar price
- 51.2%Gemini 3 Flash PreviewSimilar price
- 45.3%DeepSeek V4 Pro 0423Similar price
- 35.1%Qwen3.6 27BSimilar price
Ahead of 4 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#54 of 161
- 90.9%DeepSeek V4 Pro 0423Similar price
- 88.4%Qwen3.6 PlusNearby rank
- 88.3%
InklingNearby rank - 88.1%Mistral Large 4
- 87.9%Kimi K2.7 CodeNearby rank
- 87.9%Qwen3.7 PlusNearby rank
- 85.9%Qwen3.6 27BSimilar price
- 85.4%Nemotron 3 UltraSimilar price
- 83.3%GLM 4.7Similar price
- 51.3%Mistral Large 2.1Earlier version
+36.7 pts vs Mistral Large 2.1 (earlier version) · Ahead of 3 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#24 of 158
- 98.3%Claude Opus 4.8Nearby rank
- 98.3%DeepSeek V4.1 FlashNearby rank
- 98.3%Mistral Large 4
- 98.3%GPT-5.6 LunaNearby rank
- 98.1%Grok 4.7Nearby rank
- 96.7%DeepSeek V4 Pro 0423Similar price
- 91.1%Qwen3.6 27BSimilar price
- 86.7%Nemotron 3 UltraSimilar price
- 83.3%GLM 4.7Similar price
- 7.8%Mistral Large 2.1Earlier version
+90.6 pts vs Mistral Large 2.1 (earlier version) · Ahead of 4 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#75 of 87
- 66.8%Gemini 3 Flash PreviewSimilar price
- 47.0%DeepSeek V4 Pro 0423Similar price
- 34.3%Kimi K2.5Similar price
- 32.2%GLM 4.7Similar price
- 21.6%GPT-5 MiniNearby rank
- 20.3%Qwen3.5-FlashNearby rank
- 20.0%Mistral Large 4
- 19.1%Inkling SmallNearby rank
- 17.4%
Seed 1.6Nearby rank
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
- Arena TextArena #1291423
Level with Mistral Large 3 2512 (predecessor, overlapping 95% CIs)
- default142395% CI 1414 to 1432 · 5,742 votes
- Arena WebDevArena #471564
+331 vs Mistral Large 3 2512 (predecessor)
- default156495% CI 1549 to 1580 · 1,725 votes
- Arena VisionArena #731231
+31 vs Mistral Large 3 2512 (predecessor)
- default123195% CI 1217 to 1245 · 2,093 votes
- Agent ArenaArena #44-0.054
- default-0.05495% CI -0.075 to -0.034 · 547,357 observations
- FrontierMath-Tier-4-v2-Private#38 of 7226.8%
- high26.8%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#44 of 8758.6%
- high58.6%
Source:Epoch AI evaluation - GPQA diamond#54 of 16188.1%
+36.7 pts vs Mistral Large 2.1 (earlier version)
- high88.1%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#24 of 15898.3%
+90.6 pts vs Mistral Large 2.1 (earlier version)
- high98.3%
Source:Epoch AI evaluation - SimpleQA Verified#75 of 8720.0%
- high20.0%
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.
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.