All-time costs across the reporting window. Feeds directly into the LLM cost/quality bake-off.
| Model | Cost | Requests | Tokens (prompt+completion) | % of Spend |
|---|---|---|---|---|
| deepseek/deepseek-v4-pro | $9.65 | 411 | 64,141,608 | 37.4% |
| google/gemini-3.1-pro-preview | $8.49 | 206 | 15,306,043 | 32.9% |
| anthropic/claude-opus-4.8 | $4.77 | 58 | 2,609,382 | 18.5% |
| deepseek/deepseek-v4-flash | $2.79 | 1,495 | 44,768,708 | 10.8% |
| moonshotai/kimi-k2.6 | $0.03 | 2 | 8,737 | 0.1% |
| google/gemini-3-flash-preview | $0.03 | 11 | 19,672 | 0.1% |
| kwaipilot/kat-coder-pro-v2.5 | $0.02 | 2 | 41,302 | 0.1% |
| qwen/qwen3.7-max | $0.02 | 2 | 5,442 | 0.1% |
| google/gemini-2.5-flash | $0.01 | 2 | 3,513 | 0.0% |
| qwen/qwen3-14b | $0.00 | 1 | 862 | 0.0% |