Data scraped from OpenRouter Rankings — Programming section. Cost calculated for 2,000 input + 4,000 output tokens per task.
Generated: 2026-09-22 06:45:17 CEST
| # | Model | Author | Input / 1M | Output / 1M | Cost per task (2k in + 4k out) | Orig Rank |
|---|---|---|---|---|---|---|
| 1 | Nemotron 3 Ultra (free) | nvidia | Free | Free | Free | 7 |
| 2 | GLM 5.3 Flash | z-ai | $0.15 | $0.50 | $0.002300 | 2 |
| 3 | DeepSeek V4.1 Flash | deepseek | $0.15 | $0.60 | $0.002700 | 1 |
| 4 | MiMo-V2.5 | xiaomi | $0.14 | $0.28 | $0.001400 | 6 |
| 5 | DeepSeek V4 Flash 0731 | deepseek | $0.04 | $0.64 | $0.002640 | 4 |
| 6 | DeepSeek V4 Flash 0423 | deepseek | $0.09 | $0.18 | $0.000886 | 9 |
| 7 | Hy3 | tencent | $0.13 | $0.53 | $0.002376 | 8 |
| 8 | GPT-5.6 Luna | openai | $0.20 | $1.20 | $0.005200 | 5 |
| 9 | Hy4 preview | tencent | $0.83 | $2.50 | $0.011672 | 3 |
| 10 | GLM 5.3 | z-ai | $0.84 | $2.64 | $0.012240 | 10 |
Prices are per 1M tokens on OpenRouter. Cost = (2,000 × input/1M) + (4,000 × output/1M).