- input
- $0.20
- cached
- $0.02
- output
- $1.25
- Context
- 400Ktokens
- 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
Arena #166
- 1435
Beam 1.0 501B A23BSimilar price - 1402
Qwen3 235B-A22BNearby rank - 1402
Claude Sonnet 4Nearby rank - 1401
GPT-5.4 Nano - 1400
Qwen3 235B A22B Thinking 2507Nearby rank - 1399Qwen3 Next 80B A3B InstructNearby rank
- 1391
MiniMax M2.5Similar price - 1378
GLM 4.6VSimilar price - 1358
DeepSeek V3Similar price - 1338GPT-5 NanoPredecessor
+63 vs GPT-5 Nano (predecessor) · Ahead of 3 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 #95
- 1263DeepSeek V4.1 FlashSimilar price
- 1232MiniMax M3Similar price
- 1200
Mistral Large 3 2512Nearby rank - 1199Mistral Medium 3.5Nearby rank
- 1199GPT-5.4 Nano
- 1196
Grok 4.1 Fast (Reasoning)Nearby rank - 1193Claude Sonnet 3.7Nearby rank
- 1161GLM 4.6VSimilar price
- 1147GPT-5 NanoPredecessor
- 1145
Llama 4 MaverickSimilar price
+51 vs GPT-5 Nano (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
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
#61 of 85
- 72.9%DeepSeek V4.1 FlashSimilar price
- 42.4%Qwen3.8 27BSimilar price
- 6.5%o3Nearby rank
- 5.9%Claude Sonnet 4Nearby rank
- 5.7%GPT-5.4 Nano
- 4.9%
Gemini 2.5 ProNearby rank - 4.9%MiniMax M2.5Similar price
- 2.6%GPT-5 NanoPredecessor
- 0.0%Llama 4 MaverickSimilar price
+3.1 pts vs GPT-5 Nano (predecessor) · Ahead of 2 of 4 similarly priced models
FrontierMath-2025-02-28-Private
Best score per model · Higher is better
#19 of 58
- 27.2%GPT-5 MiniNearby rank
- 26.2%Qwen3.6 PlusNearby rank
- 25.9%GPT-5.4 Nano
- 23.1%Qwen3.6 Max PreviewNearby rank
- 21.4%
Kimi K2 ThinkingNearby rank - 10.3%Qwen3.6 FlashSimilar price
- 8.3%GPT-5 NanoPredecessor
- 4.8%Gemini 2.5 FlashSimilar price
- 1.7%DeepSeek V3Similar price
- 0.3%Mistral Medium 3Similar price
+17.6 pts vs GPT-5 Nano (predecessor) · Ahead of 4 of 4 similarly priced models
FrontierMath-Tier-4-2025-07-01-Private
Best score per model · Higher is better
#16 of 46
- 8.3%Qwen3.6 PlusSimilar price
- 8.3%Claude Sonnet 4.6Nearby rank
- 6.3%GPT-5 MiniNearby rank
- 6.3%GPT-5.4 Nano
- 4.2%Qwen3.6 Max PreviewNearby rank
- 4.2%Claude Opus 4Nearby rank
- 4.2%Gemini 2.5 FlashSimilar price
- 2.1%GLM 4.6Similar price
- 2.1%GPT-5 NanoPredecessor
- 0.0%Qwen3.6 FlashSimilar price
+4.2 pts vs GPT-5 Nano (predecessor) · Ahead of 3 of 4 similarly priced models
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#57 of 72
- 29.3%GLM 5.3Similar price
- 26.8%DeepSeek V4.1 FlashSimilar price
- 17.1%
Inkling SmallSimilar price - 12.2%Kimi K2.7 CodeNearby rank
- 12.2%GPT-5 MiniNearby rank
- 12.2%GPT-5.4 Nano
- 9.8%GPT-5.4 MiniNearby rank
- 4.9%Claude Opus 4.5Nearby rank
- 2.4%GPT-5 NanoPredecessor
- 0.0%Gemini 3.5 Flash LiteSimilar price
+9.8 pts vs GPT-5 Nano (predecessor) · Ahead of 1 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#58 of 87
- 68.8%GLM 5.3Similar price
- 67.4%DeepSeek V4.1 FlashSimilar price
- 46.3%Inkling SmallNearby rank
- 45.3%DeepSeek V4 Pro 0423Nearby rank
- 44.9%GPT-5.4 Nano
- 44.9%Grok 4.20 (Reasoning)Nearby rank
- 42.8%Grok 4.3Nearby rank
- 27.7%Gemini 3.1 Flash LiteSimilar price
- 22.5%Qwen3.6 FlashSimilar price
- 20.0%GPT-5 NanoPredecessor
+24.9 pts vs GPT-5 Nano (predecessor) · Ahead of 2 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#91 of 161
- 90.9%MiniMax M3Similar price
- 90.9%GLM 5.3Similar price
- 83.3%Qwen3.6 FlashSimilar price
- 79.2%Claude Sonnet 4Nearby rank
- 79.0%Qwen3.5-9BNearby rank
- 78.5%GPT-5.4 Nano
- 77.6%
Seed 1.6Nearby rank - 77.3%Claude Opus 4.1Nearby rank
- 71.2%DeepSeek V3Similar price
- 69.4%GPT-5 NanoPredecessor
+9.0 pts vs GPT-5 Nano (predecessor) · Ahead of 1 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#65 of 158
- 91.1%GLM 5.3Similar price
- 88.6%GPT-5.1Nearby rank
- 87.8%DeepSeek ReasonerNearby rank
- 87.8%GPT-5.4 Nano
- 86.7%Qwen3.5 PlusNearby rank
- 86.7%
Nemotron 3 UltraNearby rank - 84.4%Qwen3.6 FlashSimilar price
- 81.1%GPT-5 NanoPredecessor
- 71.1%MiniMax M3Similar price
- 48.9%DeepSeek V3Similar price
+6.7 pts vs GPT-5 Nano (predecessor) · Ahead of 3 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#83 of 87
- 42.9%DeepSeek V4.1 FlashSimilar price
- 41.0%GLM 5.3Similar price
- 19.1%Inkling SmallSimilar price
- 15.9%Qwen3.6 FlashSimilar price
- 12.7%GPT-4.1 MiniNearby rank
- 11.7%GPT-5 NanoPredecessor
- 11.7%GPT-5.4 Nano
- 10.4%Gemma 4 31BNearby rank
- 8.3%GPT-4o-mini (2024-07-18)Nearby rank
Level with GPT-5 Nano (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.
Reasoning effort
ARC-AGI-2
Score by reasoning effort · Higher is better
FrontierMath-Tiers-1-3-v2-Private
Score by reasoning effort · Higher is better
GPQA diamond
Score by reasoning effort · Higher is better
OTIS Mock AIME 2024-2025
Score by reasoning effort · Higher is better
All published scores
- Arena TextArena #1661401
+63 vs GPT-5 Nano (predecessor)
- high140195% CI 1397 to 1405 · 61,636 votes
- Arena VisionArena #951199
+51 vs GPT-5 Nano (predecessor)
- high119995% CI 1192 to 1205 · 25,407 votes
- ARC-AGI-2#61 of 855.7%
+3.1 pts vs GPT-5 Nano (predecessor)
- low1.5%
- medium1.9%
- high3.6%
- xhigh5.7%
- FrontierMath-2025-02-28-Private#19 of 5825.9%
+17.6 pts vs GPT-5 Nano (predecessor)
- high25.9%
Source:Epoch AI evaluation - FrontierMath-Tier-4-2025-07-01-Private#16 of 466.3%
+4.2 pts vs GPT-5 Nano (predecessor)
- high6.3%
Source:Epoch AI evaluation - FrontierMath-Tier-4-v2-Private#57 of 7212.2%
+9.8 pts vs GPT-5 Nano (predecessor)
- high12.2%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#58 of 8744.9%
+24.9 pts vs GPT-5 Nano (predecessor)
- none4.6%
- low20.4%
- high44.9%
Source:Epoch AI evaluation - GPQA diamond#91 of 16178.5%
+9.0 pts vs GPT-5 Nano (predecessor)
- none55.6%
- low72.2%
- high78.5%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#65 of 15887.8%
+6.7 pts vs GPT-5 Nano (predecessor)
- none46.7%
- low68.9%
- high87.8%
Source:Epoch AI evaluation - SimpleQA Verified#83 of 8711.7%
Level with GPT-5 Nano (predecessor)
- high11.7%
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.