stepfun
StepFun: Step 5 Preview
stepfun/step-5-preview
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Step 5 Preview is StepFun's flagship sparse Mixture-of-Experts model aimed at agentic work, with 600B total and 27B active parameters and a context window of up to 1M tokens, per its model card and the OpenRouter catalog. It accepts text and image input, and video too according to OpenRouter and the card's introduction, and produces text. Vendor materials highlight repository-level software engineering, long-horizon tool use, and professional knowledge work with a finance focus. Reasoning is always on, with adjustable effort.[1][2]
Strengths
- Vendor-described strength in repository-level software engineering: navigating large codebases, coordinating changes across files, and running and verifying code.[2]
- Designed for long-horizon agentic tasks with repeated tool calls, execution, verification and revision over extended context; OpenRouter exposes tool use and structured outputs.[1][2]
- Sparse MoE design activating a small fraction of its parameters per step, combined with up to 1M-token context and multimodal inputs; offered in FP8 and BF16 checkpoints with vLLM, SGLang and Transformers deployment paths.[1][2]
Limitations
- Reasoning is mandatory in the OpenRouter listing, with only effort tunable, and capability claims rest mainly on the vendor card; the pack contains no independent evaluations, so validate on your own workload.[1][2]
- Self-hosting is demanding: vendor examples require 4 to 16 datacenter GPUs (B300/H200), and the BF16 file-shard generation workflow is not yet validated end to end.[2]
- Sources disagree on video input (the card's spec table lists only text and image), the license is a custom other whose terms are not summarized here, and the knowledge cutoff is unstated.[1][2]
Where to use it
- Worth evaluating for agentic coding assistants that inspect, edit and verify large repositories.[2]
- A candidate for long-document and financial analysis workflows combining quantitative reasoning with charts, screenshots and reports.[2]
- Consider for multimodal, long-context assistants where text, image and possibly video inputs feed extended reasoning and tool use; confirm video support on your endpoint.[1][2]
Sources
- [1] OpenRouter catalog · StepFun: Step 5 PreviewSources checked: Nov 13, 2026
- [2] Model documentation · StepFun: Step 5 PreviewSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts architecture (27B active / 600B total parameters). It performs strongly in software engineering and professional...
Specifications
- Model ID
- stepfun/step-5-preview
- Canonical ID
- stepfun/step-5-preview-20261008
- Creator
- stepfun
- Added to OpenRouter
- 08.10.2026
- Context length
- 1,000,000
- Modality
- text+image+video->text
- Input modalities
- text, image, video
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $1 / out $2.7
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | BB — next 30% of the cohort | reasoning: mandatory · default medium · efforts: high, medium, low · parameters: include_reasoning, reasoning, reasoning_effort |
| Coding | No measurement | — |
| Agents | No measurement | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: mandatory · default medium · efforts: high, medium, low · parameters: include_reasoning, reasoning, reasoning_effort |
| Document analysis | No measurement | — |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | input: video |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Intelligence index | 28.5 | BB — next 30% of the cohort | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| Featherlessfp4 | 262,144 | $1 | $2.7 | — | — | — |
| StepFunfp8 | 1,000,000 | $1 | $2.7 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.