Lab · China
StepFun
AI models from StepFun: 43 models in the catalog, version timeline, prices and sourced benchmarks.
- 43 models in the catalog
- 37 with open weights
- 6 closed or without public weights
- 0 with sourced benchmarks
- 3 with a page here
We count grouped models: Instruct, Base, quantized and size variants of the same line become one. Behind that there are 54 Hugging Face repositories, 3 OpenRouter endpoints and 11 historical records at Epoch AI.
Version timeline by family
Main model lines
Research and others (6)
This group holds benchmarks, datasets, research artifacts, and lines from brands that are not in this area's catalog.
Tap a point to see the version.Tap or hover over a point to see the version. Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch. Family and version were grouped by rule from the name.
Version prices over time (US$ per 1M tokens, log scale) ?what it is
- input
- output
Source: OpenRouter, latest version we checked. Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch.
Models with a page (3)
| Model | Family | Weights | Use cases | Input US$/1M | Context | GPQA | Dates in sources |
|---|---|---|---|---|---|---|---|
| Step 5 Preview | Step | open | Text, Agents | US$ 1.00 | 1,000,000 | no data | 2026 |
| Step 3.7 Flash | Step Flash | open | Text, Agents | US$ 0.20 | 262,144 | no data | 2026 |
| Step 3.5 Flash | Step Flash | open | Text, Agents | US$ 0.10 | 262,144 | no data | 2026 |
Other lab models in the catalog (16)
| Model | Modality | Weights | Context | Input US$/1M | Downloads (30 days) | 1ª |
|---|---|---|---|---|---|---|
| GOT-OCR2_0 | multimodal | open | no data | no data | 229,669 | 2024 |
| Step-3.5-Flash | texto | open | 262,144 | US$ 0.10 | 189,811 | 2026 |
| step3 | multimodal | open | no data | no data | 45,265 | 2025 |
| Step-3.7-Flash | multimodal | open | 262,144 | US$ 0.20 | 21,807 | 2026 |
| stepvideo-t2v | imagem, vídeo | open | no data | no data | 94 | 2025 |
| Step-Audio | multimodal | open | no data | no data | 261 | 2025 |
| Step-Audio-2-mini | multimodal, sem rótulo | open | no data | no data | 4,328 | 2025 |
| Step1X-Edit | imagem | open | no data | no data | 204 | 2025 |
| Step-5 | multimodal, sem rótulo | open | 1,000,000 | US$ 1.00 | 10,894 | 2026 |
| Step-Audio-R1.1 | multimodal | open | no data | no data | 651 | 2026 |
| Step-Audio-TTS | voz | open | no data | no data | 237 | 2025 |
| Step-Audio-EditX | sem rótulo | open | no data | no data | 2,389 | 2025 |
| GELab-Zero | multimodal | open | no data | no data | 1,355 | 2025 |
| Step-3.5-Flash-GGUF-Q4_K_S | texto | open | no data | no data | 1,017 | 2026 |
| Step-Audio-R1 | multimodal | open | no data | no data | 118 | 2025 |
| Step1X-3D | imagem | open | no data | no data | no data | 2025 |
Sources
- Hugging Face Hub, metadados públicos do Hub; licença de cada modelo no repositório, latest version we checked.
- OpenRouter, dados públicos da API; exports de benchmarks e rankings em CC BY 4.0, latest version we checked.
- Epoch AI, CC BY 4.0, latest version we checked.
Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch. Family and version were grouped by rule from the name.