Calculator
What a task costs, not what a token costs
Illustrative token mixes for comparison. They are not measured traces from a real coding session or chat. The default below is a coding-agent session. Change the preset to recompute in the browser. With JavaScript off, the coding-session table still stands.
3 million input tokens, 90% served from cache, 60 thousand output tokens. Cache-write fees are not included.
| Model | Session | vs Opus | vs Sol | Cache price |
|---|---|---|---|---|
| qwen3.8-flash | $0.116 | 96% cheaper | 92% cheaper | published |
| glm-5.3-flash | $0.096 | 97% cheaper | 94% cheaper | published |
| hy3 | $0.160 | 95% cheaper | 89% cheaper | published |
| deepseek-v4.1-flash | $0.066 | 98% cheaper | 95% cheaper | published |
| deepseek-v4-flash | $0.087 | 97% cheaper | 94% cheaper | published |
| minimax-m3 | $0.324 | 89% cheaper | 78% cheaper | published |
| mimo-v2.6-pro | $0.192 | 93% cheaper | 87% cheaper | published |
| mimo-v2.6-flash | $0.066 | 98% cheaper | 95% cheaper | published |
| deepseek-v4-pro | $0.616 | 79% cheaper | 58% cheaper | published |
| glm-5.3 | $0.522 | 82% cheaper | 64% cheaper | published |
| kimi-k3 | $1.35 | 54% cheaper | 8% cheaper | published |
| qwen3.8-max-prime | $3.27 | 11% dearer | 122% dearer | published |
| step-3.5-flash | $0.318 | 89% cheaper | 78% cheaper | not listed |
| Anthropic: Claude Opus 5.5 | $2.94 | — | — | published |
| OpenAI: GPT-6.1 Sol | $1.47 | — | — | published |
The percentages in this table compare session cost. The price table's percentages compare blended token price. They answer different questions.