- Context
- 1.1Mtokens
- Released
- 1mo 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 #98
- 1453
GPT-5.6 LunaPredecessor - 1452
Gemma 4 31BSimilar price - 1448
MiMo-V2.6-FlashSimilar price - 1441
Qwen3.5 397B A17BNearby rank - 1441
GLM 4.7Nearby rank - 1441GPT-6 Luna
- 1437Gemma 4 26B A4BNearby rank
- 1437GPT-5.2Nearby rank
- 1422
Claude Haiku 5.5Similar price - 1326Qwen3 30B A3BSimilar price
-12 vs GPT-5.6 Luna (predecessor) · Ahead of 2 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #45
- 1638MiMo-V2.6-FlashSimilar price
- 1609
Muse EmberSimilar price - 1594Claude Haiku 5.5Similar price
- 1582
DeepSeek V4 Flash 0423Nearby rank - 1581DeepSeek V4 Pro 0813Nearby rank
- 1577GPT-6 Luna
- 1573
Step 5 PreviewNearby rank - 1564
Mistral Large 4Nearby rank - 1519GPT-5.6 LunaPredecessor
- 1370Gemma 4 31BSimilar price
+58 vs GPT-5.6 Luna (predecessor) · Ahead of 1 of 4 similarly priced models
Arena Vision
Best Arena rating per model · Higher is better
Arena #68
- 1265Gemma 4 31BSimilar price
- 1255GPT-5.6 LunaPredecessor
- 1248MiMo-V2.6-FlashSimilar price
- 1242Gemma 4 26B A4BNearby rank
- 1242
Grok 4.3Nearby rank - 1235GPT-6 Luna
- 1235Gemini 3.1 Flash Lite PreviewNearby rank
- 1234MiMo-V2.5Nearby rank
- 1128Llama 4 ScoutSimilar price
- 1019
Nova Lite 1.0Similar price
-20 vs GPT-5.6 Luna (predecessor) · Ahead of 2 of 4 similarly priced models
Arena Search
Best Arena rating per model · Higher is better
Arena #12
- 1213Claude Opus 5Nearby rank
- 1213Claude Opus 5.5Nearby rank
- 1211GPT-6 Luna
- 1210Claude Opus 4.7Nearby rank
- 1205Grok 4.5Nearby rank
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
Arena #26
- +0.020DeepSeek V4.1 Flash 1017Similar price
- +0.018Muse WonderNearby rank
- +0.017Muse EmberSimilar price
- +0.016GPT-6 Luna
- +0.010GLM 5.3Nearby rank
- +0.006
Kimi K3Nearby rank - +0.005Claude Haiku 5.5Similar price
- -0.000MiMo-V2.6-FlashSimilar price
Ahead of 1 of 4 similarly priced models
ARC-AGI-2
Best score per model · Higher is better
#39 of 85
- 78.1%DeepSeek V4.1 Flash 1017Similar price
- 61.3%Claude Haiku 5.5Similar price
- 60.4%Kimi K3Nearby rank
- 59.5%GPT-5.6 LunaPredecessor
- 59.3%GPT-6 Luna
- 54.2%GPT-5.2 ProNearby rank
- 52.9%GPT-5.2Nearby rank
- 1.3%Qwen3 235B A22B Instruct 2507Similar price
- 0.0%Llama 4 ScoutSimilar price
-0.2 pts vs GPT-5.6 Luna (predecessor) · Ahead of 2 of 4 similarly priced models
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#20 of 72
- 61.0%GPT-5.6 LunaPredecessor
- 58.5%GPT-5.4 ProNearby rank
- 56.1%Claude Opus 4.8Nearby rank
- 56.1%GPT-6 Luna
- 49.0%GPT-5.4Nearby rank
- 46.3%
Qwen3.8 MaxNearby rank - 46.3%Claude Haiku 5.5Similar price
- 46.3%DeepSeek V4.1 Flash 1017Similar price
- 24.4%DeepSeek V4 Flash 0731Similar price
- 2.4%GPT-5 NanoSimilar price
-4.9 pts vs GPT-5.6 Luna (predecessor) · Ahead of 4 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#20 of 87
- 82.1%GPT-5.6 LunaPredecessor
- 80.0%Claude Opus 4.8Nearby rank
- 78.9%GPT-6 Luna
- 78.6%GPT-5.4Nearby rank
- 75.1%Claude Haiku 5.5Similar price
- 71.2%DeepSeek V4.1 Flash 1017Similar price
- 20.4%Qwen3.6 35B A3BSimilar price
- 18.2%Qwen3.5-FlashSimilar price
-3.2 pts vs GPT-5.6 Luna (predecessor) · Ahead of 4 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#42 of 161
- 93.2%DeepSeek V4.1 Flash 1017Similar price
- 91.6%GPT-5.6 LunaPredecessor
- 90.5%Claude Opus 4.6Nearby rank
- 90.5%Claude Sonnet 5Nearby rank
- 90.5%GPT-6 Luna
- 90.2%Claude Opus 4.7Nearby rank
- 90.2%GLM 5.3 FlashNearby rank
- 89.6%Claude Haiku 5.5Similar price
- 75.8%Gemma 4 31BSimilar price
- 61.7%Qwen3 30B A3BSimilar price
-1.1 pts vs GPT-5.6 Luna (predecessor) · Ahead of 3 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#19 of 158
- 99.4%DeepSeek V4.1 Flash 1017Similar price
- 98.9%Claude Haiku 5.5Similar price
- 98.9%Claude Opus 5Nearby rank
- 98.9%Gemini 3.8 FlashNearby rank
- 98.9%GPT-6 Luna
- 98.6%DeepSeek V4 Pro 0813Nearby rank
- 98.3%Claude Opus 4.8Nearby rank
- 98.3%GPT-5.6 LunaPredecessor
- 73.3%Gemma 4 31BSimilar price
- 62.8%Qwen3 30B A3BSimilar price
+0.6 pts vs GPT-5.6 Luna (predecessor) · Ahead of 2 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#49 of 87
- 42.9%DeepSeek V4.1 FlashNearby rank
- 42.5%DeepSeek V4.1 Flash 1017Similar price
- 41.4%GPT-6 Luna
- 41.0%GPT-5.6 LunaPredecessor
- 41.0%GLM 5.3Nearby rank
- 23.8%Claude Haiku 5.5Similar price
- 20.3%Qwen3.5-FlashSimilar price
- 10.4%Gemma 4 31BSimilar price
+0.4 pts vs GPT-5.6 Luna (predecessor) · Ahead of 3 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
Agent Arena
Score by reasoning effort · Higher is better
ARC-AGI-2
Score by reasoning effort · Higher is better
All published scores
- Arena TextArena #981441
-12 vs GPT-5.6 Luna (predecessor)
- max144195% CI 1437 to 1445 · 107,021 votes
- Arena WebDevArena #451577
+58 vs GPT-5.6 Luna (predecessor)
- max157795% CI 1572 to 1583 · 30,498 votes
- Arena VisionArena #681235
-20 vs GPT-5.6 Luna (predecessor)
- max123595% CI 1226 to 1244 · 6,202 votes
- Arena SearchArena #121211
- max121195% CI 1204 to 1218 · 46,776 votes
- Agent ArenaArena #26+0.016
- low-0.06095% CI -0.067 to -0.053 · 3,812,346 observations
- medium-0.02195% CI -0.025 to -0.016 · 11,530,251 observations
- max+0.01695% CI +0.011 to +0.020 · 42,848,344 observations
- ARC-AGI-2#39 of 8559.3%
-0.2 pts vs GPT-5.6 Luna (predecessor)
- none0.0%
- low4.6%
- medium18.1%
- high31.4%
- xhigh41.9%
- max59.3%
- FrontierMath-Tier-4-v2-Private#20 of 7256.1%
-4.9 pts vs GPT-5.6 Luna (predecessor)
- max56.1%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#20 of 8778.9%
-3.2 pts vs GPT-5.6 Luna (predecessor)
- max78.9%
Source:Epoch AI evaluation - GPQA diamond#42 of 16190.5%
-1.1 pts vs GPT-5.6 Luna (predecessor)
- max90.5%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#19 of 15898.9%
+0.6 pts vs GPT-5.6 Luna (predecessor)
- max98.9%
Source:Epoch AI evaluation - SimpleQA Verified#49 of 8741.4%
+0.4 pts vs GPT-5.6 Luna (predecessor)
- max41.4%
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.