StepFun
8 storiesStepFun is a Chinese AI company that develops the Step family of large models. Recent coverage centers on Step-5-Preview, a 600B-parameter flagship with 1M context aimed at agentic and professional work, offered free for a week in Cline and OpenCode. It reached #2 on OpenRouter's weekly leaderboard with 19.5T tokens used, scored 44 on the Artificial Analysis Intelligence Index at $0.71 per task, and ranked #7 among open-weight models on the Vals leaderboard.
Related topics
A Reddit user shared receiving a 451 censorship error from a US inference provider, with content blocked by StepFun CN, while merely making a presentation about cryptographic algorithms. The experience strengthened their belief in local LLMs and reminded that cloud use is subject to provider terms.
One week since launch, Step 5 Preview is already #2 on OpenRouter's weekly leaderboard, with 19.5T tokens used.
It has also held #2 on the daily leaderboard since Oct 10.
Thanks to everyone building with us.
StepFun launched Step-5-Preview, a 600B-parameter flagship model with 1M context for agentic work, delivering frontier-level performance in software engineering and professional knowledge work, with particular strength in finance.
Step 5 Preview is now on @OpenRouter's New & Trending leaderboard.
Thanks to everyone giving it a try, and to our partners bringing it into your workflow.
Step 5 Preview is now free in @cline for one week.
Our flagship model for agentic and professional work.
Fire up Cline and show us what you build!
[Quoting @cline]: Step 5 Preview is now free in Cline.
On DeepSWE it scores ahead of Kimi K3 and GLM-5.3, making…
Your next build is on us.
Step 5 Preview is now free in @opencode for one week. Our flagship model for agentic and professional work.
Keep your workflow. Switch your model.
[Quoting @opencode]: Step 5 Preview is now free on OpenCode for a week
- 1M Context
-…
Step 5 Preview just landed on the Vals leaderboard. 🚀
[Quoting @ValsAI]: StepFun debuts on the Vals leaderboard with Step 5 Preview, ranking #7 among open-weight models, just ahead of Qwen 3.8 Max. It costs $2.54/task.
Step 5 Preview pushes our intelligence–cost Pareto frontier outward.
It comes in at 44 on the Artificial Analysis Intelligence Index, $0.71 per task.
Thanks @ArtificialAnlys for putting Step 5 Preview through the full evaluation.
More to come on Oct 15.
[Quot…
That is everything