| Model | Input / output per 1M | Log probabilities | Quantization |
|---|---|---|---|
| GLM 5.3 | $0.783 / $2.46 | no | fp8 |
| DeepSeek V4 Pro 0423 | $1.60 / $3.20 | yes | fp8 |
| gpt-oss-120b | $0.05 / $0.25 | yes | fp4 |
| gpt-oss-20b | $0.04 / $0.15 | yes | fp4 |
| Gemma 4 31B | $0.14 / $0.40 | yes | bf16 |
| DeepSeek V4.1 Flash | $0.195 / $0.78 | yes | fp8 |
| DeepSeek V4 Flash 0731 | $0.409 / $1.23 | yes | fp8 |
| GLM 5.3 Flash | $0.113 / $0.375 | no | fp8 |
| Hy4 preview | $0.834 / $2.50 | no | fp8 |
| Qwen3.8 27B | $0.42 / $3.00 | yes | unknown |
| MiniMax M3 | $0.30 / $1.20 | no | fp8 |
| Kimi K2.6 | $0.80 / $3.40 | no | unknown |
| Llama 4 Maverick | $0.27 / $0.85 | no | fp8 |
| Llama 3.3 70B Instruct | $0.135 / $0.40 | yes | bf16 |