- input
- $0.10
- output
- $0.30
- Context
- 1.3Mtokens
- Released
- 1y ago
- Size
- 109Bparameters
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 #264
- 1393
Step 3.5 FlashSimilar price - 1382
Qwen3 30B A3B Instruct 2507Similar price - 1365
Gemma 3 27BSimilar price - 1364
GLM 4.7 FlashSimilar price - 1325
Claude Haiku 3.5Nearby rank - 1322
GPT-4.1 NanoNearby rank - 1322
Llama 4 Scout - 1321
Molmo 2 8BNearby rank - 1318Llama 3.3 70B InstructPredecessor
Level with Llama 3.3 70B Instruct (predecessor, overlapping 95% CIs) · Ahead of 0 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 #127
- 1248
MiMo-V2.6-FlashSimilar price - 1160Gemma 3 27BSimilar price
- 1147GPT-5 NanoSimilar price
- 1145Llama 4 MaverickNearby rank
- 1140
Mistral Small 3.2Nearby rank - 1128Llama 4 Scout
- 1127Claude Haiku 3.5Nearby rank
- 1127Mistral Small 3.1 24BNearby rank
- 1019
Nova Lite 1.0Similar price
Ahead of 1 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
Not ranked on Agent Arena yet
ARC-AGI-2
Best score per model · Higher is better
#78 of 85
- 61.4%
DeepSeek V4 Flash 0731Similar price - 61.3%Claude Haiku 5.5Similar price
- 2.6%GPT-5 NanoSimilar price
- 0.0%Gemini 2.5 Pro Preview 06-05Nearby rank
- 0.0%Llama 4 MaverickNearby rank
- 0.0%Llama 4 Scout
- 0.0%Magistral SmallNearby rank
- 0.0%GPT-4.1 MiniNearby rank
- 0.0%GPT-4.1 NanoSimilar price
Ahead of 0 of 4 similarly priced models
FrontierMath-2025-02-28-Private
Best score per model · Higher is better
#58 of 58
- 8.3%GPT-5 NanoSimilar price
- 6.2%Qwen3.5-FlashSimilar price
- 1.0%GPT-4.1 NanoSimilar price
- 0.3%GPT-4o (2024-08-06)Nearby rank
- 0.3%GPT-4o (2024-11-20)Nearby rank
- 0.0%Llama 4 Scout
Ahead of 0 of 3 similarly priced models
GPQA diamond
Best score per model · Higher is better
#132 of 161
- 69.4%GPT-5 NanoSimilar price
- 63.8%Qwen3 14BSimilar price
- 60.5%GLM 4.7 FlashSimilar price
- 55.6%Qwen3 30B A3B Instruct 2507Nearby rank
- 55.3%Claude Sonnet 3.5 v2Nearby rank
- 51.8%Llama 4 Scout
- 51.3%Mistral Large 2.1Nearby rank
- 49.2%GPT-4o (2024-08-06)Nearby rank
- 47.7%Gemma 3 27BSimilar price
- 47.4%Llama 3.3 70B InstructPredecessor
+4.4 pts vs Llama 3.3 70B Instruct (predecessor) · Ahead of 1 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#139 of 158
- 81.1%GPT-5 NanoSimilar price
- 66.4%Qwen3 14BSimilar price
- 58.3%GLM 4.7 FlashSimilar price
- 22.5%Gemma 3 27BSimilar price
- 8.5%Mistral Large 2407Nearby rank
- 8.1%Qwen2.5 72B InstructNearby rank
- 7.8%Llama 4 Scout
- 7.8%Mistral Large 2.1Nearby rank
- 7.5%Gemma 3 4BNearby rank
- 5.1%Llama 3.3 70B InstructPredecessor
+2.6 pts vs Llama 3.3 70B Instruct (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 #2641322
Level with Llama 3.3 70B Instruct (predecessor, overlapping 95% CIs)
- default132295% CI 1317 to 1326 · 29,777 votes
- Arena VisionArena #1271128
- default112895% CI 1118 to 1137 · 6,431 votes
- ARC-AGI-2#78 of 850.0%
- default0.0%
- FrontierMath-2025-02-28-Private#58 of 580.0%
- default0.0%
Source:Epoch AI evaluation - GPQA diamond#132 of 16151.8%
+4.4 pts vs Llama 3.3 70B Instruct (predecessor)
- default51.8%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#139 of 1587.8%
+2.6 pts vs Llama 3.3 70B Instruct (predecessor)
- default7.8%
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.