CoreWeave
Price history
Blended ($/Mtok, 75% input / 25% output) price over time for every model this host serves. Click a legend entry to hide or show its line.
Hosted offerings
| Model | Variant | Input $/Mtok | Output $/Mtok | Confidence |
|---|---|---|---|---|
| OpenAI: gpt-oss-120b | fp4 | $0.030000 | $0.170000 | high |
| OpenAI: gpt-oss-20b | fp4 | $0.030000 | $0.130000 | high |
| NVIDIA: Nemotron 3.5 Lightning (free) | – | $0.070000 | $0.200000 | high |
| IBM: Granite 4.2 8B | – | $0.100000 | $0.150000 | high |
| Google: Gemma 4 26B A4B (free) | – | $0.100000 | $0.300000 | high |
| Google: Gemma 4 31B (free) | fp4 | $0.100000 | $0.340000 | high |
| DeepSeek: DeepSeek V4 Flash 0731 | fp8 | $0.130000 | $0.280000 | high |
| Z.ai: GLM 5.3 Flash | nvfp4 | $0.150000 | $0.500000 | high |
| DeepSeek: DeepSeek V4.1 Flash | fp8 | $0.200000 | $0.650000 | high |
| Meta: Llama 3.1 8B Instruct | – | $0.220000 | $0.220000 | high |
| MiniMax: MiniMax M3 | fp4 | $0.230000 | $0.960000 | high |
| Qwen: Qwen3.6 35B A3B | fp8 | $0.250000 | $1.250000 | high |
| Qwen: Qwen3.8 27B (free) | fp8 | $0.400000 | $3.000000 | high |
| DeepSeek: DeepSeek V3.1 | fp8 | $0.550000 | $1.650000 | high |
| MoonshotAI: Kimi K2.6 | fp4 | $0.650000 | $3.410000 | high |
| MoonshotAI: Kimi K2.7 Code | int4 | $0.710000 | $3.500000 | high |
| Meta: Llama 3.3 70B Instruct | – | $0.710000 | $0.710000 | high |
| Z.ai: GLM 5.2 | fp4 | $0.760000 | $2.420000 | high |
| DeepSeek: DeepSeek V4 Pro 0813 | fp8 | $1.310000 | $3.960000 | high |