Vendor · Language models · Image models · Video models
StepFun has 8 models in the directory. The strongest is Step 5 Preview: Rank 27 to 36 of 95 in the overall index.
In the directory: 4 language models, 1 image model and 3 video models. 7 of them with open weights that can run locally.
Every model there ever was. Models with a profile lead to every benchmark, price and source.
| Step 3.7 Flash | May 2026 | Open source | n/a | $0.20 / $1.15 | 108 | 256K | 41 |
| Step 3.5 Flash 2603 | Apr 2, 2026 | Active | n/a | n/a | 151 | 256K | 18 |
| Step 3.5 Flash | February 2026 | Open source | n/a | $0.10 / $0.30 | 160 | 256K | 19 |
| Step 5 Preview | n/a | Preview | 43Rank 27 to 36 of 95 | $1 / $2.70 | 182 | n/a | 2 |
| NextStep-1.1 | Dec 24, 2025 | Active | n/a | n/a |
| Step-Video-TI2V 30B | Mar 17, 2025 | Active | n/a | n/a |
| Step-Video-T2V 30B | Feb 17, 2025 | Active | n/a | n/a |
| Step-Video-T2V Turbo | n/a | Active | n/a | n/a |
Language model · May 2026
New in Step 3.7 Flash is image understanding in StepFun's fast Step line. The sparse MoE accepts text and images, returns text, activates about 11 billion parameters, and combines a 256K context with selectable low, medium, or high reasoning effort.
Language model · Apr 2, 2026
Step 3.5 Flash 2603 is StepFun’s large sparse Mixture-of-Experts foundation model for reasoning, coding, and agentic work. Its official card lists 196.81 billion total parameters and about 11 billion active per token. MTP-3 and a 256,000-token context are intended to combine fast generation with long workflows. This checkpoint is for text-to-text open Apache-2.0 weights, not a multimodal Step configuration.
Language model · February 2026
Step 3.5 Flash combines roughly 197 billion total parameters with about 11 billion active and a 256K context using sliding-window attention. The fast open-weights model processes text; StepFun also evaluates a Parallel Thinking variant.
Image model · Dec 24, 2025
NextStep-1.1 is an image model from StepFun. The reviewed model specifications list 15B parameters.
Video model · Mar 17, 2025
Step-Video-TI2V 30B is a video model from StepFun. Step-Video-TI2V 30B handles text and images in one video model; the model description lists up to 102 frames.
Video model · Feb 17, 2025
Step-Video-T2V 30B is a video model from StepFun. Step-Video-T2V 30B is a text-to-video model.
Video model · n/a
Step-Video-T2V Turbo is a video model from StepFun. Step-Video-T2V Turbo supports text-to-video.
Derived from the overall index, the API price and the measured speed of the current models.
The strongest choice
Step 5 Preview
The vendor's highest overall index
The fastest
Step 3.5 Flash
160 tokens/s, a typical model manages 98
The 20 best language models, with the models from StepFun highlighted. “All” and the filter show every model in the index.
20 of 95 models
Provider
Weights
Price (blended per 1M tokens)
Released
Anthropic
OpenAI
Meta
Moonshot AI
DeepSeek
xAI
Z.ai
Xiaomi
Alibaba
StepFun
Mistral AI
MiniMax
Thinking Machines Lab
Inception
| Vendor | Models | Model types | Best language model | Top 10 | Open weights | New in 2026 |
|---|---|---|---|---|---|---|
| Anthropic | 33 | 33 language models | Claude Opus 5.5 (Rank 1 to 3) | 4 | 0 | 13 |
| OpenAI | 65 | 46 language models, 6 image models, 2 video models, 9 audio models, 2 specialist models | GPT-6 Astra (Rank 4 to 8) | 2 | 4 | 27 |
| 67 | 43 language models, 5 image models, 6 video models, 11 audio models, 2 specialist models | Gemini 4 Carbon (Rank 1 to 5) | 2 | 19 | 26 | |
| Meta | 32 | 29 language models, 1 image model, 1 video model, 1 specialist model | Muse Wonder Full (Rank 6 to 9) | 2 | 26 | 7 |
| DeepSeek | 25 | 23 language models, 2 image models | DeepSeek V4.1 Flash 1017 (Rank 13 to 20) | 0 | 25 | 9 |
| Moonshot AI | 12 | 12 language models | Kimi K3.1 Pro (Rank 10 to 13) | 0 | 10 | 7 |
| xAI | 18 | 12 language models, 3 image models, 3 video models | Grok 4.7 (Rank 18 to 24) | 0 | 1 | 12 |
| StepFun | 8 | 4 language models, 1 image model, 3 video models | Step 5 Preview (Rank 27 to 36) | 0 | 7 | 3 |
By release date. Grey: retired models.
3 releases since January 2026, by release date.
Prices per 1M tokens for the current language models, according to StepFun or the provider linked in each model profile.
| Current model | Input | Output | Context |
|---|---|---|---|
| Step 5 PreviewPreview | $1 | $2.70 | n/a |
| Step 3.5 Flash | $0.10 | $0.30 | 256K |
| Step 3.7 Flash | $0.20 | $1.15 | 256K |
The directory lists 8 StepFun models: 4 language models, 1 image model and 3 video models. It counts every model StepFun ever released, including retired ones. 7 are currently available, the newest being Step 3.7 Flash, Step 3.5 Flash 2603 and Step 3.5 Flash.
Step 5 Preview. It is at rank 27 to 36 of 95 in the overall index, which combines independent benchmarks from several sources.
Step 3.7 Flash, released in May 2026. The timeline below shows every release by date.
Through the API per 1M tokens (input / output): Step 3.5 Flash $0.10 / $0.30 and Step 5 Preview $1 / $2.70. The prices of all 3 current language models are in the price table.
In the overall index, no. The best OpenAI model, GPT-6 Astra (rank 4 to 8 of 95), is ahead of Step 5 Preview (rank 27 to 36 of 95). Single tasks can differ.
Yes, 7 of the 8 models have open weights. Each model profile lists how much memory it needs.
Step 5 Preview with 182 tokens per second according to Artificial Analysis. A typical model manages 98.
Up to 256,000 tokens (≈ 380 A4 pages), which 3 of the current models reach, including Step 3.7 Flash. Every model's value is in the model list.
Data as of Nov 2, 2026