- Context
- 1Mtokens
- Released
- 3mo 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
#23 of 28
- 74.1%
- 73.8%
- 73.6%
- 72.7%
- 69.9%
- 69.6%
- 69.0%
- 68.5%
- 67.5%
- 67.2%
- 46.7%Gemini 3.6 Flash
Arena Text
Best Arena rating per model · Higher is better
Arena #25
- 1484
Qwen3.8 MaxNearby rank - 1483Kimi K3.1 ProNearby rank
- 1483Gemini 3.6 Flash
- 1483GLM 5.3Nearby rank
- 1482GPT-6.1 SolNearby rank
- 1478Gemini 3.5 FlashPredecessor
- 1457GLM 5Similar price
- 1448
Step 5 PreviewSimilar price - 1442Grok 4.3Similar price
- 1290
Nova Pro 1.0Similar price
Level with Gemini 3.5 Flash (predecessor, overlapping 95% CIs) · Ahead of 4 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #55
- 1573Step 5 PreviewSimilar price
- 1541
Muse Spark 1.1Nearby rank - 1539Claude Sonnet 5Nearby rank
- 1538Gemini 3.6 Flash
- 1534Muse Spark 1.2Nearby rank
- 1526Muse Spark OpenNearby rank
- 1500Gemini 3.5 FlashPredecessor
- 1435GLM 5Similar price
- 1400
Qwen3.5 397B A17BSimilar price - 1357Grok 4.3Similar price
+38 vs Gemini 3.5 Flash (predecessor) · Ahead of 3 of 4 similarly priced models
Arena Vision
Best Arena rating per model · Higher is better
Arena #28
- 1281Gemini 3.5 FlashPredecessor
- 1280Grok 4.5Nearby rank
- 1280Gemini 3.1 Pro PreviewNearby rank
- 1280Gemini 3.6 Flash
- 1279Muse Spark 1.1Nearby rank
- 1278GPT-5.2 ChatNearby rank
- 1246Qwen3.5 397B A17BSimilar price
- 1245Step 5 PreviewSimilar price
- 1242Grok 4.3Similar price
- 1019Nova Pro 1.0Similar price
Level with Gemini 3.5 Flash (predecessor, overlapping 95% CIs) · Ahead of 4 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
Arena #30
- 1480Muse Spark 1.1Similar price
- 1459Gemini 3.1 Pro PreviewNearby rank
- 1459Gemini 3.5 FlashPredecessor
- 1454Gemini 3.6 Flash
- 1452Gemma 4 31BNearby rank
- 1451Gemini 3 Pro PreviewNearby rank
- 1450Kimi K2.6Similar price
- 1419Claude Haiku 4.5Similar price
- 1406GLM 5V TurboSimilar price
Level with Gemini 3.5 Flash (predecessor, overlapping 95% CIs) · Ahead of 2 of 4 similarly priced models
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
#36 of 85
- 72.1%Gemini 3.5 FlashPredecessor
- 65.1%Grok 4.20Similar price
- 61.3%
DeepSeek V4 Pro 0813Nearby rank - 60.4%Claude Sonnet 4.6Nearby rank
- 60.4%Gemini 3.6 Flash
- 60.4%Kimi K3Nearby rank
- 59.5%GPT-5.6 LunaNearby rank
- 36.5%
InklingSimilar price - 18.9%GPT-5.4 MiniSimilar price
- 4.9%GLM 5Similar price
-11.7 pts vs Gemini 3.5 Flash (predecessor) · Ahead of 3 of 4 similarly priced models
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#47 of 72
- 29.3%GLM 5.2Similar price
- 26.8%Gemini 3.5 FlashPredecessor
- 25.6%Kimi K2.6Similar price
- 24.4%DeepSeek V4 Flash 0731Nearby rank
- 24.4%Grok 4.5Nearby rank
- 22.0%Gemini 3.6 Flash
- 22.0%Gemini 3.8 FlashNearby rank
- 22.0%GPT-5Nearby rank
- 14.6%Grok 4.3Similar price
- 9.8%GPT-5.4 MiniSimilar price
-4.9 pts vs Gemini 3.5 Flash (predecessor) · Ahead of 2 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#43 of 87
- 62.8%Gemini 3.5 FlashPredecessor
- 59.6%Gemini 3.1 Pro PreviewNearby rank
- 59.2%GLM 5.2Nearby rank
- 58.9%Gemini 3.6 Flash
- 58.6%
Mistral Large 4Nearby rank - 57.5%DeepSeek V4 Flash 0731Nearby rank
- 57.2%Kimi K2.6Similar price
- 51.2%GPT-5.4 MiniSimilar price
- 42.8%Grok 4.3Similar price
- 31.2%Qwen3.5 397B A17BSimilar price
-3.9 pts vs Gemini 3.5 Flash (predecessor) · Ahead of 4 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#12 of 161
- 94.3%GPT-6 SolNearby rank
- 94.3%GPT-6.1 TerraNearby rank
- 94.1%Gemini 3.6 Flash
- 94.0%Grok 4.6Nearby rank
- 93.9%Claude Fable 5.2Nearby rank
- 92.8%Gemini 3.5 FlashPredecessor
- 90.8%Kimi K2.6Similar price
- 88.8%Grok 4.3Similar price
- 87.8%GLM 5Similar price
- 86.4%Qwen3.5 397B A17BSimilar price
+1.3 pts vs Gemini 3.5 Flash (predecessor) · Ahead of 4 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#45 of 158
- 96.1%Kimi K2.6Similar price
- 95.6%Gemini 3.5 FlashPredecessor
- 94.4%Claude Opus 4.6Nearby rank
- 94.4%DeepSeek V4 Flash 0731Nearby rank
- 94.2%Gemini 3.6 Flash
- 93.9%GLM 5.3 FlashNearby rank
- 93.3%Qwen3.6 PlusNearby rank
- 93.3%Grok 4.3Similar price
- 88.9%Qwen3.5 397B A17BSimilar price
- 80.0%GLM 5Similar price
-1.4 pts vs Gemini 3.5 Flash (predecessor) · Ahead of 3 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#14 of 87
- 66.8%Gemini 3 Flash PreviewNearby rank
- 66.2%Gemini 3.5 FlashPredecessor
- 66.2%Gemini 3.6 Flash
- 63.0%GPT-5.5Nearby rank
- 60.7%GPT-6 SolNearby rank
- 34.9%Kimi K2.6Similar price
- 34.2%GLM 5.2Similar price
- 33.2%Grok 4.3Similar price
- 29.4%GPT-5.4 MiniSimilar price
Level with Gemini 3.5 Flash (predecessor) · Ahead of 4 of 4 similarly priced models
DeepSWE 1.1
Primary coding-agent benchmark · 113 tasks · Data generated 2026-09-22
- pass@4
- 75.2%
- Average cost / attempt
- $4.42
- Benchmark runs
- 4
pass@1 is the success rate across scored attempts; pass@4 is the share of tasks solved in at least one of four attempts. Costs are reported by DeepSWE for each configuration. These results are not directly comparable to evaluations using different harnesses. Official source JSON.
Reasoning effort
ARC-AGI-2
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 #251483
Level with Gemini 3.5 Flash (predecessor, overlapping 95% CIs)
- high148395% CI 1478 to 1487 · 35,405 votes
- Arena WebDevArena #551538
+38 vs Gemini 3.5 Flash (predecessor)
- high153895% CI 1530 to 1545 · 8,524 votes
- Arena VisionArena #281280
Level with Gemini 3.5 Flash (predecessor, overlapping 95% CIs)
- high128095% CI 1272 to 1287 · 10,352 votes
- Arena DocumentArena #301454
Level with Gemini 3.5 Flash (predecessor, overlapping 95% CIs)
- high145495% CI 1443 to 1465 · 3,031 votes
- ARC-AGI-2#36 of 8560.4%
-11.7 pts vs Gemini 3.5 Flash (predecessor)
- minimal2.6%
- low30.4%
- medium50.4%
- high60.4%
- FrontierMath-Tier-4-v2-Private#47 of 7222.0%
-4.9 pts vs Gemini 3.5 Flash (predecessor)
- high22.0%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#43 of 8758.9%
-3.9 pts vs Gemini 3.5 Flash (predecessor)
- high58.9%
Source:Epoch AI evaluation - GPQA diamond#12 of 16194.1%
+1.3 pts vs Gemini 3.5 Flash (predecessor)
- minimal85.9%
- low86.4%
- high94.1%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#45 of 15894.2%
-1.4 pts vs Gemini 3.5 Flash (predecessor)
- minimal80.0%
- low82.2%
- high94.2%
Source:Epoch AI evaluation - SimpleQA Verified#14 of 8766.2%
Level with Gemini 3.5 Flash (predecessor)
- high66.2%
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.