Every model at this usage
Monthly per-token cost, cheapest first. Standard Compute is flat at $9–$399 regardless of volume.
| Model | Per-token / mo | vs Standard ($39) |
|---|
| Qwen3 235B A22B · Alibaba (Qwen) | $50 | 1.3× more |
| DeepSeek V4 Flash · DeepSeek | $57 | 1.5× more |
| Llama 4 Maverick · Meta | $122 | 3.1× more |
| DeepSeek V3.2 · DeepSeek | $134 | 3.4× more |
| MiniMax M2.1 · MiniMax | $216 | 5.5× more |
| Codestral · Mistral | $216 | 5.5× more |
| MiniMax M3 · MiniMax | $243 | 6.2× more |
| Gemini 3.1 Flash Lite · Google | $247 | 6.3× more |
| Qwen3.7 Plus · Alibaba (Qwen) | $259 | 6.6× more |
| DeepSeek V4 Pro · DeepSeek | $274 | 7.0× more |
| GLM 4.7 · Z.AI | $338 | 8.7× more |
| GLM 4.6 · Z.AI | $350 | 9.0× more |
| Mistral Large 3 · Mistral | $360 | 9.2× more |
| Devstral 2 · Mistral | $360 | 9.2× more |
| DeepSeek R1 (0528) · DeepSeek | $418 | 11× more |
| Kimi K2.7 Code · Moonshot AI | $552 | 14× more |
| Kimi K2.6 · Moonshot AI | $604 | 15× more |
| GLM 5.2 · Z.AI | $718 | 18× more |
| GPT-5.4 Mini · OpenAI | $743 | 19× more |
| Grok 4.3 · xAI | $788 | 20× more |
| Grok 4.20 · xAI | $788 | 20× more |
| Claude Haiku 4.5 · Anthropic | $900 | 23× more |
| Qwen3.7 Max · Alibaba (Qwen) | $900 | 23× more |
| Gemini 3.5 Flash · Google | $1,485 | 38× more |
| Gemini 3.1 Pro · Google | $1,980 | 51× more |
| GPT-5.3-Codex · OpenAI | $2,048 | 53× more |
| GPT-5.4 · OpenAI | $2,475 | 63× more |
| Claude Sonnet 4.6 · Anthropic | $2,700 | 69× more |
| Claude Opus 4.8 · Anthropic | $4,500 | 115× more |
| GPT-5.5 · OpenAI | $4,950 | 127× more |
Estimates use public OpenRouter per-token pricing (input + output) at the usage above. Real bills vary with caching, retries, and tool calls — which agents do constantly, so per-token tends to run higher than estimated. Standard Compute pricing is flat and unlimited within fair use.