- input
- $0.40
- cached
- $0.10
- output
- $1.60
- Context
- 1Mtokens
- 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
Arena #192
- 1416
Qwen3.5-122B-A10BSimilar price - 1415
Mistral Large 3 2512Similar price - 1387
Solar Pro 4Nearby rank - 1387Mistral Medium 3Nearby rank
- 1383
GPT-4.1 Mini - 1382Qwen3 30B A3B Instruct 2507Nearby rank
- 1379
Gemini 2.5 Flash-Lite Preview 09-2025Nearby rank - 1318
Llama 3.3 70B InstructSimilar price - 1318GPT-4o-mini (2024-07-18)Predecessor
- 1290Gemma 2 27BSimilar price
+65 vs GPT-4o-mini (2024-07-18) (predecessor) · Ahead of 2 of 4 similarly priced models
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
Arena #92
- 1234
MiMo-V2.5Similar price - 1226Qwen3.5-122B-A10BSimilar price
- 1208
Inkling SmallSimilar price - 1206
Claude Sonnet 4Nearby rank - 1201GPT-4.1 Mini
- 1200Mistral Large 3 2512Similar price
- 1199Mistral Medium 3.5Nearby rank
- 1097GPT-4o-mini (2024-07-18)Predecessor
+104 vs GPT-4o-mini (2024-07-18) (predecessor) · Ahead of 0 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
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
- 42.4%Qwen3.8 27BSimilar price
- 40.1%Inkling SmallSimilar price
- 10.3%Gemini 3.5 Flash LiteSimilar price
- 1.1%
R1 0528Similar price - 0.0%Llama 4 ScoutNearby rank
- 0.0%Magistral SmallNearby rank
- 0.0%GPT-4.1 Mini
- 0.0%GPT-4.1 NanoNearby rank
- 0.0%GPT-4o (2024-11-20)Nearby rank
- 0.0%GPT-4o-mini (2024-07-18)Predecessor
Level with GPT-4o-mini (2024-07-18) (predecessor) · Ahead of 0 of 4 similarly priced models
FrontierMath-2025-02-28-Private
Best score per model · Higher is better
#42 of 58
- 26.2%Qwen3.6 PlusSimilar price
- 4.8%Gemini 2.5 FlashSimilar price
- 4.5%Claude Opus 4Nearby rank
- 4.5%GPT-4.1 Mini
- 4.1%Claude Sonnet 3.7Nearby rank
- 4.1%Claude Sonnet 4Nearby rank
- 3.8%
GLM 4.6Similar price - 0.3%Mistral Medium 3Similar price
Ahead of 2 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#84 of 87
- 67.4%DeepSeek V4.1 FlashSimilar price
- 46.3%Inkling SmallSimilar price
- 38.2%Qwen3.6 PlusSimilar price
- 26.0%Gemini 3.5 Flash LiteSimilar price
- 18.2%Qwen3.5-FlashNearby rank
- 12.6%Claude Opus 4.1Nearby rank
- 6.7%GPT-4.1 Mini
- 6.0%GPT-4.1Nearby rank
- 0.7%GPT-4o-mini (2024-07-18)Predecessor
+6.0 pts vs GPT-4o-mini (2024-07-18) (predecessor) · Ahead of 0 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#114 of 161
- 88.5%Inkling SmallSimilar price
- 88.4%Qwen3.6 PlusSimilar price
- 67.6%DeepSeek V3 0324Nearby rank
- 66.9%GPT-4.1Nearby rank
- 65.8%GPT-4.1 Mini
- 65.7%Qwen3 32BNearby rank
- 65.4%
QwQ PlusNearby rank - 47.4%Llama 3.3 70B InstructSimilar price
- 37.7%GPT-4o-mini (2024-07-18)Predecessor
- 36.5%Gemma 2 27BSimilar price
+28.1 pts vs GPT-4o-mini (2024-07-18) (predecessor) · Ahead of 2 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#122 of 158
- 93.3%Qwen3.6 PlusSimilar price
- 90.0%Inkling SmallSimilar price
- 55.6%DeepSeek-R1-Distill-Qwen-32BNearby rank
- 48.9%DeepSeek V3Nearby rank
- 44.7%GPT-4.1 Mini
- 43.9%DeepSeek R1 0528 Qwen3 8BNearby rank
- 38.3%GPT-4.1Nearby rank
- 6.9%GPT-4o-mini (2024-07-18)Predecessor
- 5.1%Llama 3.3 70B InstructSimilar price
- 1.4%Gemma 2 27BSimilar price
+37.8 pts vs GPT-4o-mini (2024-07-18) (predecessor) · Ahead of 2 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#82 of 87
- 44.1%Qwen3.6 PlusSimilar price
- 42.9%DeepSeek V4.1 FlashSimilar price
- 34.3%
Kimi K2.5Similar price - 19.1%Inkling SmallSimilar price
- 15.9%Qwen3.6 FlashNearby rank
- 13.2%Claude Haiku 4.5Nearby rank
- 12.7%GPT-4.1 Mini
- 11.7%GPT-5 NanoNearby rank
- 11.7%GPT-5.4 NanoNearby rank
- 8.3%GPT-4o-mini (2024-07-18)Predecessor
+4.4 pts vs GPT-4o-mini (2024-07-18) (predecessor) · 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 #1921383
+65 vs GPT-4o-mini (2024-07-18) (predecessor)
- default138395% CI 1379 to 1387 · 38,657 votes
- Arena VisionArena #921201
+104 vs GPT-4o-mini (2024-07-18) (predecessor)
- default120195% CI 1193 to 1208 · 40,154 votes
- ARC-AGI-2#78 of 850.0%
Level with GPT-4o-mini (2024-07-18) (predecessor)
- default0.0%
- FrontierMath-2025-02-28-Private#42 of 584.5%
- default4.5%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#84 of 876.7%
+6.0 pts vs GPT-4o-mini (2024-07-18) (predecessor)
- default6.7%
Source:Epoch AI evaluation - GPQA diamond#114 of 16165.8%
+28.1 pts vs GPT-4o-mini (2024-07-18) (predecessor)
- default65.8%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#122 of 15844.7%
+37.8 pts vs GPT-4o-mini (2024-07-18) (predecessor)
- default44.7%
Source:Epoch AI evaluation - SimpleQA Verified#82 of 8712.7%
+4.4 pts vs GPT-4o-mini (2024-07-18) (predecessor)
- default12.7%
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.