- input
- $1.40
- cached
- $0.26
- output
- $4.40
- Context
- 205Ktokens
- Released
- 7mo ago
- Size
- 754Bparameters
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 #56
- 1496
Muse Spark 1.3Similar price - 1483
Kimi K3.1 ProSimilar price - 1467
MiMo-V2.5-ProNearby rank - 1466
Grok 4.5Nearby rank - 1465
GLM 5.1 - 1464
DeepSeek V4 Pro 0813Nearby rank - 1461
Claude Sonnet 5Nearby rank - 1457GLM 5Predecessor
- 1443
InklingSimilar price - 1414Claude Haiku 4.5Similar price
+8 vs GLM 5 (predecessor) · Ahead of 2 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #67
- 1679
GPT-6.1 TerraSimilar price - 1655Muse Spark 1.3Similar price
- 1515
Qwen3.7 MaxSimilar price - 1512GPT-5.5Nearby rank
- 1509Kimi K2.6Nearby rank
- 1508GLM 5.1
- 1506
Hy3Nearby rank - 1500
Gemini 3.5 FlashNearby rank - 1435GLM 5Predecessor
- 1330Claude Haiku 4.5Similar price
+73 vs GLM 5 (predecessor) · 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
FrontierMath-2025-02-28-Private
Best score per model · Higher is better
#12 of 58
- 39.0%Gemini 3.5 FlashSimilar price
- 39.0%Kimi K2.6Similar price
- 36.9%Gemini 3.1 Pro PreviewNearby rank
- 35.6%Gemini 3 Flash PreviewNearby rank
- 33.4%GLM 5.1
- 32.4%GPT-5Nearby rank
- 32.4%Claude Sonnet 4.6Nearby rank
- 28.3%GPT-5.4 MiniSimilar price
- 16.4%GLM 5Predecessor
- 5.9%Claude Haiku 4.5Similar price
+17.0 pts vs GLM 5 (predecessor) · Ahead of 2 of 4 similarly priced models
FrontierMath-Tier-4-2025-07-01-Private
Best score per model · Higher is better
#11 of 46
- 14.6%Gemini 3.5 FlashSimilar price
- 14.6%Kimi K2.6Similar price
- 12.5%GPT-5Nearby rank
- 12.5%GPT-5.1Nearby rank
- 12.5%GLM 5.1
- 8.3%Qwen3.6 PlusNearby rank
- 8.3%Claude Sonnet 4.6Nearby rank
- 2.1%GLM 5Predecessor
- 2.1%Claude Haiku 4.5Similar price
- 2.1%GPT-5.4 MiniSimilar price
+10.4 pts vs GLM 5 (predecessor) · Ahead of 2 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#62 of 87
- 90.9%GPT-6.1 TerraSimilar price
- 74.4%Muse Spark 1.3Similar price
- 64.6%Qwen3.7 MaxSimilar price
- 42.8%Grok 4.3Nearby rank
- 38.2%Qwen3.6 PlusNearby rank
- 36.8%GLM 5.1
- 35.1%Qwen3.6 27BNearby rank
- 34.4%Claude Opus 4.5Nearby rank
- 33.3%InklingSimilar price
Ahead of 1 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#45 of 161
- 94.3%GPT-6.1 TerraSimilar price
- 90.9%Qwen3.7 MaxSimilar price
- 90.2%Claude Opus 4.7Nearby rank
- 90.2%GLM 5.3 FlashNearby rank
- 89.9%GLM 5.1
- 89.8%DeepSeek V4.1 FlashNearby rank
- 89.6%Claude Haiku 5.5Nearby rank
- 88.3%InklingSimilar price
- 87.8%GLM 5Predecessor
- 71.2%Claude Haiku 4.5Similar price
+2.1 pts vs GLM 5 (predecessor) · Ahead of 2 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#47 of 158
- 100.0%GPT-6.1 TerraSimilar price
- 99.2%Muse Spark 1.3Similar price
- 95.6%Qwen3.7 MaxSimilar price
- 93.3%Qwen3.7 PlusNearby rank
- 93.3%Grok 4.3Nearby rank
- 93.3%GLM 5.1
- 92.2%Grok 4.20 (Reasoning)Nearby rank
- 91.4%Gemini 3 Pro PreviewNearby rank
- 80.0%GLM 5Predecessor
- 66.7%Claude Haiku 4.5Similar price
+13.3 pts vs GLM 5 (predecessor) · Ahead of 1 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#60 of 87
- 60.3%Muse Spark 1.2Similar price
- 57.3%GPT-6.1 TerraSimilar price
- 55.8%Qwen3.7 MaxSimilar price
- 34.3%Kimi K2.5Nearby rank
- 34.2%GLM 5.2Nearby rank
- 34.0%GLM 5.1
- 33.7%Claude Sonnet 5Nearby rank
- 33.6%DeepSeek V4 Flash 0731Nearby rank
- 13.2%Claude Haiku 4.5Similar price
Ahead of 1 of 4 similarly priced models
SWE-bench Verified
Best score per model · Higher is better
#17 of 35
- 79.3%Gemini 3.5 FlashSimilar price
- 77.3%Qwen3.7 MaxSimilar price
- 76.7%Kimi K2.6Similar price
- 75.2%Claude Sonnet 4.6Nearby rank
- 74.8%GPT-5.3-CodexNearby rank
- 74.2%GLM 5.1
- 73.8%Kimi K2.5Nearby rank
- 73.8%GPT-5.2Nearby rank
- 73.6%GPT-5Similar price
- 72.1%GLM 5Predecessor
+2.1 pts vs GLM 5 (predecessor) · Ahead of 1 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 #561465
+8 vs GLM 5 (predecessor)
- default146595% CI 1462 to 1469 · 67,074 votes
- Arena WebDevArena #671508
+73 vs GLM 5 (predecessor)
- default150895% CI 1501 to 1515 · 10,531 votes
- FrontierMath-2025-02-28-Private#12 of 5833.4%
+17.0 pts vs GLM 5 (predecessor)
- default33.4%
Source:Epoch AI evaluation - FrontierMath-Tier-4-2025-07-01-Private#11 of 4612.5%
+10.4 pts vs GLM 5 (predecessor)
- default12.5%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#62 of 8736.8%
- none24.9%
- default36.8%
Source:Epoch AI evaluation - GPQA diamond#45 of 16189.9%
+2.1 pts vs GLM 5 (predecessor)
- default89.9%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#47 of 15893.3%
+13.3 pts vs GLM 5 (predecessor)
- default93.3%
Source:Epoch AI evaluation - SimpleQA Verified#60 of 8734.0%
- default34.0%
Source:Epoch AI evaluation - SWE-bench Verified#17 of 3574.2%
+2.1 pts vs GLM 5 (predecessor)
- default74.2%
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.