Best models
DeepSeek, GLM, and Kimi ranked on reasoning, agentic coding, and automation benchmarks.
| # | Model | Avg | Price /1M |
|---|---|---|---|
| 1 | DeepSeek V4.1-Flash 0910 | 64.2 | $0.750 |
| 2 | Kimi K3 | 62.7 | — |
| 3 | GLM 5.3 | 61.5 | — |
| 4 | DeepSeek V4-Pro 0813 | 60.3 | $2.64 |
| 5 | DeepSeek V4.1-Flash 1017 | 60.0 | $0.750 |
| 6 | DeepSeek V4-Flash 0731 | 55.0 | — |
Scored on: GPQA Diamond, HLE, DeepSWE v1.1, Automation-Bench · See the full leaderboard