- input
- $2.00
- cached
- $0.30
- output
- $6.00
- Context
- 500Ktokens
- Released
- 4mo 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
#19 of 28
- 74.1%
- 73.8%
- 73.6%
- 72.7%
- 69.9%
- 69.6%
- 69.0%
- 68.5%
- 67.5%
- 67.2%
- 53.8%Grok 4.5
Arena Text
Best Arena rating per model · Higher is better
Arena #54
- 1483Kimi K3.1 ProSimilar price
- 1478Gemini 3.5 FlashSimilar price
- 1467
ERNIE 5.1Nearby rank - 1467
MiMo-V2.5-ProNearby rank - 1466Grok 4.5
- 1465GLM 5.1Nearby rank
- 1464
DeepSeek V4 Pro 0813Nearby rank - 1455GPT-5.1Similar price
- 1442Grok 4.3Predecessor
- 1427
Mistral Medium 3.5Similar price
+24 vs Grok 4.3 (predecessor) · Ahead of 2 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #51
- 1676
Qwen3.8 Max (0902)Similar price - 1557Claude Opus 4.7Nearby rank
- 1555Claude Opus 4.8Nearby rank
- 1552Grok 4.5
- 1547Claude Opus 4.6Nearby rank
- 1541
Muse Spark 1.1Nearby rank - 1500Gemini 3.5 FlashSimilar price
- 1417GPT-5Similar price
- 1357Grok 4.3Predecessor
- 1263Mistral Medium 3.5Similar price
+195 vs Grok 4.3 (predecessor) · Ahead of 3 of 4 similarly priced models
Arena Vision
Best Arena rating per model · Higher is better
Arena #26
- 1282GPT-5.4Nearby rank
- 1281Gemini 3.5 FlashSimilar price
- 1280Grok 4.5
- 1280Gemini 3.1 Pro PreviewNearby rank
- 1280Gemini 3.6 FlashNearby rank
- 1263Claude Sonnet 5.5Similar price
- 1251GPT-5.1Similar price
- 1242Grok 4.3Predecessor
- 1199Mistral Medium 3.5Similar price
+39 vs Grok 4.3 (predecessor) · Ahead of 3 of 4 similarly priced models
Arena Search
Best Arena rating per model · Higher is better
Arena #16
- 1248GPT-6.1 SolSimilar price
- 1229GPT-6 SolSimilar price
- 1211GPT-6 LunaNearby rank
- 1210Claude Opus 4.7Nearby rank
- 1205Grok 4.5
- 1201Gemini 3 Pro PreviewNearby rank
- 1194Claude Opus 4.8Nearby rank
- 1194Grok 4.3Predecessor
- 1142
Sonar ProSimilar price - 1138Sonar Reasoning ProSimilar price
+12 vs Grok 4.3 (predecessor) · Ahead of 2 of 4 similarly priced models
Arena Document
Best Arena rating per model · Higher is better
Arena #24
- 1480Muse Spark 1.1Similar price
- 1469Claude Sonnet 5Similar price
- 1468Gemini 3.7 FlashNearby rank
- 1467Claude Opus 4.5Nearby rank
- 1466Grok 4.5
- 1460GPT-5.6 LunaNearby rank
- 1460GPT-5.5 InstantNearby rank
- 1459Gemini 3.5 FlashSimilar price
- 1416GPT-5.1Similar price
Ahead of 1 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
#42 of 85
- 83.9%Claude Sonnet 5.5Similar price
- 72.1%Gemini 3.5 FlashSimilar price
- 65.1%Grok 4.20Earlier version
- 60.4%Kimi K3Similar price
- 54.2%GPT-5.2 ProNearby rank
- 52.9%GPT-5.2Nearby rank
- 52.6%Grok 4.5
- 42.4%Qwen3.8 27BNearby rank
- 40.1%
Inkling SmallNearby rank - 17.6%GPT-5.1Similar price
-12.5 pts vs Grok 4.20 (earlier version) · Ahead of 1 of 4 similarly priced models
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#45 of 72
- 80.5%Claude Sonnet 5.5Similar price
- 34.1%Qwen3.8 Max (0902)Similar price
- 26.8%Gemini 3.5 FlashSimilar price
- 25.6%Kimi K2.6Nearby rank
- 24.4%DeepSeek V4 Flash 0731Nearby rank
- 24.4%Grok 4.5
- 22.0%Gemini 3.6 FlashNearby rank
- 22.0%Gemini 3.8 FlashNearby rank
- 22.0%GPT-5Similar price
- 14.6%Grok 4.3Predecessor
+9.8 pts vs Grok 4.3 (predecessor) · Ahead of 1 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#46 of 87
- 88.8%Claude Sonnet 5.5Similar price
- 65.6%Qwen3.8 Max (0902)Similar price
- 62.8%Gemini 3.5 FlashSimilar price
- 57.5%DeepSeek V4 Flash 0731Nearby rank
- 57.2%Kimi K2.6Nearby rank
- 57.2%Grok 4.5
- 55.8%GPT-5 ProNearby rank
- 55.8%GLM 5.3 FlashNearby rank
- 55.4%GPT-5Similar price
- 42.8%Grok 4.3Predecessor
+14.4 pts vs Grok 4.3 (predecessor) · Ahead of 1 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#17 of 161
- 93.9%Claude Opus 5Nearby rank
- 93.5%GPT-5.6 SolNearby rank
- 93.4%Grok 4.5
- 93.3%GPT-5.6 TerraNearby rank
- 93.3%GPT-5.4Nearby rank
- 92.8%Gemini 3.5 FlashSimilar price
- 92.3%Qwen3.8 Max (0902)Similar price
- 88.8%Grok 4.3Predecessor
- 87.6%GPT-5.1Similar price
- 49.0%Mistral Large 2407Similar price
+4.6 pts vs Grok 4.3 (predecessor) · Ahead of 4 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#30 of 158
- 100.0%Qwen3.8 Max (0902)Similar price
- 97.8%Claude Opus 4.7Nearby rank
- 97.8%GPT-5.4Nearby rank
- 97.8%Grok 4.5
- 97.2%Gemini 3.7 FlashNearby rank
- 97.2%Kimi K3Nearby rank
- 95.6%Gemini 3.5 FlashSimilar price
- 93.3%Grok 4.3Predecessor
- 91.4%GPT-5Similar price
- 8.5%Mistral Large 2407Similar price
+4.4 pts vs Grok 4.3 (predecessor) · Ahead of 3 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#34 of 87
- 66.2%Gemini 3.5 FlashSimilar price
- 50.1%GPT-5Similar price
- 49.3%Grok 4.6Nearby rank
- 48.7%
Qwen3 MaxNearby rank - 48.3%Grok 4.5
- 48.0%GPT-5.1Nearby rank
- 47.3%Qwen3.8 Max (0902)Similar price
- 46.5%Claude Sonnet 5.5Similar price
- 33.2%Grok 4.3Predecessor
+15.1 pts vs Grok 4.3 (predecessor) · Ahead of 2 of 4 similarly priced models
DeepSWE 1.1
Primary coding-agent benchmark · 113 tasks · Data generated 2026-09-22
- pass@4
- 77.9%
- Average cost / attempt
- $2.42
- 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
All published scores
- Arena TextArena #541466
+24 vs Grok 4.3 (predecessor)
- default146695% CI 1461 to 1470 · 39,155 votes
- Arena WebDevArena #511552
+195 vs Grok 4.3 (predecessor)
- default155295% CI 1546 to 1559 · 10,800 votes
- Arena VisionArena #261280
+39 vs Grok 4.3 (predecessor)
- default128095% CI 1274 to 1287 · 13,991 votes
- Arena SearchArena #161205
+12 vs Grok 4.3 (predecessor)
- default120595% CI 1200 to 1211 · 105,282 votes
- Arena DocumentArena #241466
- default146695% CI 1459 to 1474 · 7,447 votes
- ARC-AGI-2#42 of 8552.6%
-12.5 pts vs Grok 4.20 (earlier version)
- low33.1%
- medium52.6%
- high52.6%
- FrontierMath-Tier-4-v2-Private#45 of 7224.4%
+9.8 pts vs Grok 4.3 (predecessor)
- high24.4%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#46 of 8757.2%
+14.4 pts vs Grok 4.3 (predecessor)
- high57.2%
Source:Epoch AI evaluation - GPQA diamond#17 of 16193.4%
+4.6 pts vs Grok 4.3 (predecessor)
- high93.4%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#30 of 15897.8%
+4.4 pts vs Grok 4.3 (predecessor)
- high97.8%
Source:Epoch AI evaluation - SimpleQA Verified#34 of 8748.3%
+15.1 pts vs Grok 4.3 (predecessor)
- high48.3%
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.