- input
- $0.60
- cached
- $0.11
- output
- $2.20
- Context
- 205Ktokens
- Released
- 10mo ago
- Size
- 358Bparameters
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 #97
- 1450
Kimi K2.5Similar price - 1442
Grok 4.3Nearby rank - 1441
Qwen3.5 397B A17BNearby rank - 1441
GLM 4.7 - 1441
GPT-6 LunaNearby rank - 1437
Gemma 4 26B A4BNearby rank - 1426
Nemotron 3 UltraSimilar price - 1424GLM 4.6Predecessor
- 1423
Mistral Large 4Similar price - 1421
R1 0528Similar price
+17 vs GLM 4.6 (predecessor) · Ahead of 3 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #93
- 1632
Hy4 previewSimilar price - 1564Mistral Large 4Similar price
- 1448DeepSeek V4 Pro 0423Similar price
- 1436Kimi K2.5Similar price
- 1435Gemini 3.5 Flash LiteNearby rank
- 1435GLM 5Nearby rank
- 1435GLM 4.7
- 1434
MiMo-V2-ProNearby rank - 1420GPT-5.2Nearby rank
- 1342GLM 4.6Predecessor
+93 vs GLM 4.6 (predecessor) · Ahead of 0 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
#47 of 58
- 35.6%Gemini 3 Flash PreviewSimilar price
- 21.4%Kimi K2 ThinkingSimilar price
- 8.5%Qwen3 235B A22B Thinking 2507Similar price
- 4.1%
Claude Sonnet 4Nearby rank - 3.8%GLM 4.6Predecessor
- 2.4%GLM 4.7
- 2.1%Claude Sonnet 3.5 v2Nearby rank
- 1.7%DeepSeek V3Nearby rank
- 0.3%Mistral Medium 3Similar price
-1.4 pts vs GLM 4.6 (predecessor) · Ahead of 1 of 4 similarly priced models
FrontierMath-Tier-4-2025-07-01-Private
Best score per model · Higher is better
#39 of 46
- 6.3%GPT-5 MiniSimilar price
- 4.2%Gemini 3 Flash PreviewSimilar price
- 2.1%GLM 4.6Predecessor
- 0.0%Kimi K2 ThinkingSimilar price
- 0.0%Qwen3 235B A22B Thinking 2507Similar price
- 0.0%Qwen3.5-FlashNearby rank
- 0.0%Qwen3.6 FlashNearby rank
- 0.0%GLM 4.7
-2.1 pts vs GLM 4.6 (predecessor) · Ahead of 0 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#77 of 161
- 88.1%Mistral Large 4Similar price
- 85.9%Qwen3.6 27BSimilar price
- 85.4%Nemotron 3 UltraSimilar price
- 83.3%Gemini 3.5 Flash LiteNearby rank
- 83.3%Qwen3.6 FlashNearby rank
- 83.3%GLM 4.7
- 82.5%GPT-5.5 InstantNearby rank
- 82.3%Claude Sonnet 4.5Nearby rank
- 76.3%R1 0528Similar price
Ahead of 1 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#81 of 158
- 98.3%Mistral Large 4Similar price
- 91.1%Qwen3.6 27BSimilar price
- 86.7%Nemotron 3 UltraSimilar price
- 84.4%Qwen3.5-FlashNearby rank
- 84.4%Qwen3.6 FlashNearby rank
- 83.3%GLM 4.7
- 83.1%Kimi K2 Thinking TurboNearby rank
- 82.2%Gemma 4 26B A4BNearby rank
- 66.4%R1 0528Similar price
Ahead of 1 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#64 of 87
- 66.8%Gemini 3 Flash PreviewSimilar price
- 47.0%DeepSeek V4 Pro 0423Similar price
- 34.3%Kimi K2.5Similar price
- 33.6%DeepSeek V4 Flash 0731Nearby rank
- 33.2%Grok 4.3Nearby rank
- 32.2%GLM 4.7
- 31.1%GPT-4.1Nearby rank
- 30.7%Claude Sonnet 4.5Nearby rank
- 20.0%Mistral Large 4Similar price
Ahead of 1 of 4 similarly priced models
Terminal Bench
Best score per model · Higher is better
#36 of 49
- 64.3%Gemini 3 Flash PreviewSimilar price
- 61.6%GPT-5.1-Codex-MiniSimilar price
- 57.3%Grok 4.20Similar price
- 43.2%Kimi K2.5Similar price
- 35.5%Claude Haiku 4.5Nearby rank
- 34.8%GPT-5 MiniNearby rank
- 33.4%GLM 4.7
- 32.6%Gemini 2.5 ProNearby rank
- 30.0%
MiniMax-M2Nearby rank - 24.5%GLM 4.6Predecessor
+8.9 pts vs GLM 4.6 (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.
All published scores
- Arena TextArena #971441
+17 vs GLM 4.6 (predecessor)
- default144195% CI 1435 to 1447 · 12,167 votes
- Arena WebDevArena #931435
+93 vs GLM 4.6 (predecessor)
- default143595% CI 1423 to 1447 · 3,975 votes
- FrontierMath-2025-02-28-Private#47 of 582.4%
-1.4 pts vs GLM 4.6 (predecessor)
- default2.4%
Source:Epoch AI evaluation - FrontierMath-Tier-4-2025-07-01-Private#39 of 460.0%
-2.1 pts vs GLM 4.6 (predecessor)
- default0.0%
Source:Epoch AI evaluation - GPQA diamond#77 of 16183.3%
- default83.3%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#81 of 15883.3%
- default83.3%
Source:Epoch AI evaluation - SimpleQA Verified#64 of 8732.2%
- default32.2%
Source:Epoch AI evaluation - Terminal Bench#36 of 4933.4%
+8.9 pts vs GLM 4.6 (predecessor)
- default33.4%Terminus 2
- default33.3%Crux
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.