stepfun-ai models
4 stepfun-ai models from StepFun on Hugging Face, from 561M to 321.0B parameters, published by StepFun in BF16, with contexts from 32K to 256K tokens. The smallest official model, GOT-OCR-2.0-hf, needs about 1.3 GB of VRAM at its published precision; the cheapest live fit is RTX 3070 at $0.088/hr.
Pick a size
One row per official stepfun-ai size: the VRAM it needs at each precision, the cheapest GPU that holds it today, and the KV cache for a 32K-token context.
| Model | Parameters | Native VRAM | FP8 VRAM | INT4 VRAM | Live GPU fit (native) | Est. $/hr | KV cache at 32K |
|---|---|---|---|---|---|---|---|
| GOT-OCR-2.0-hf | 561M | 1.3 GB | 0.6 GB | 0.3 GB | RTX 3070 | $0.088/hr | 3.0 GB |
| GOT-OCR2_0 | 716M | 1.6 GB | 0.8 GB | 0.4 GB | RTX 3070 | $0.088/hr | 3.0 GB |
| Step-3.5-Flash | 199.4B | 446 GB | 223 GB | 111 GB | RTX PRO 6000 × 5 | $6.88/hr | not published for this architecture |
| step3 | 321.0B | 717 GB | 359 GB | 179 GB | AMD MI300X × 4 | $10.52/hr | 122.0 GB |
VRAM is weights times a flat 1.2 overhead; the KV cache is a separate per-model figure at 16-bit, shown where the architecture is published. See the methodology.
Official models (4)
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| GOT-OCR-2.0-hf | 561M | BF16 | 1.3 GB | RTX 3070 | $0.088/hr |
| GOT-OCR2_0 | 716M | BF16 | 1.6 GB | RTX 3070 | $0.088/hr |
| Step-3.5-Flash | 199.4B | BF16 | 446 GB | RTX PRO 6000 × 5 | $6.88/hr |
| step3 | 321.0B | BF16 | 717 GB | AMD MI300X × 4 | $10.52/hr |
VRAM is for the precision the model is published in: the weights times a flat 1.2 overhead, with the KV cache not included. See the methodology. The fit shown is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does.