models & pricing
pricing is usage-based, in USD, billed per 1M tokens. we offer the following models for training and serverless inference.
- training is metered on the tokens your training run processes: rollout generation plus weight updates.
- inference pricing is the same for base model inference and serving trained checkpoints. other models are included for use within our synthetic generation pipeline and judge rewards.
training
per 1M training tokens| model | size | context | training cost |
|---|---|---|---|
| Qwen3-VL 4B Instruct Qwen/Qwen3-VL-4B-Instruct | small 4.4b | 256k | $0.10
launch price
|
| Qwen3.5 4B Qwen/Qwen3.5-4B | small 4.7b | 256k | $0.10
launch price
|
| Qwen3.5 35B-A3B Qwen/Qwen3.5-35B-A3B | medium 36b | 256k | $0.20
launch price
|
| Gemma 4 26B-A4B google/gemma-4-26B-A4B-it | medium 26b | 256k | $0.20
launch price
|
| Step 3.7 Flash stepfun-ai/Step-3.7-Flash | large 198b | 256k | contact us |
| Hy3 tencent/Hy3 | large 295b | 256k | contact us |
need a different model? get in touch with us.
inference
per 1M input / output tokens| model | input | output |
|---|---|---|
| GPT-5.6 Sol gpt-5.6-sol | $5.00 | $30.00 |
| GPT-5.6 Luna gpt-5.6-luna | $1.00 | $6.00 |
| GPT-5.6 Terra gpt-5.6-terra | $2.50 | $15.00 |
| GPT-5.4 gpt-5.4 | $2.50 | $15.00 |
| GPT-5.4 Mini gpt-5.4-mini | $0.75 | $4.50 |
| GPT-5.4 Nano gpt-5.4-nano | $0.20 | $1.25 |
| Grok 4.1 Fast Non-Reasoning grok-4-1-fast-non-reasoning | $0.20 | $0.50 |
| Grok 4.3 grok-4.3 | $1.25 | $2.50 |
| Qwen3.5 4B qwen3.5-4b | $0.03 | $0.15 |
| Qwen3.5 35B-A3B qwen3.5-35b-a3b | $0.25 | $2.00 |
| Gemma 4 26B-A4B gemma-4-26b-a4b-it | $0.06 | $0.33 |
embeddings
| model | input |
|---|---|
| Text Embedding 3 Large text-embedding-3-large | $0.143 |