The Model Index
Models/StepFun

Step 5 Preview

StepFunClosed weightsFlagship

Step 5 Preview is a closed AI model from StepFun with an index score of 163 (GPT-4 = 100), provisional.

163145–180Index scoreProvisional · one third-party tester so far; ranked once a second confirms it
$1.0/ $2.7Price per million tokensinput / output · via OpenRouter
1MContext windowtokens it reads at once
18 Sep 2026Released2 months ago

Where it sitson the frontier

801001201401601802002023202420252026GPT-4 at launch = 100Step 5 Preview

Claude Opus 4.6 reached this score 7 months earlier.

Test resultsby area · tick = best by any model

Reasoning & knowledgeno evidence yet
GPQA Diamond (Epoch AI)
no result
Humanity's Last Exam
no result
Mathsevidence
FrontierMath (tiers 1-3)
no result
Codingevidence
SWE-bench Verified
no result
Terminal-Bench 4.0
32%
Agentsevidence
APEX-Agents
no result
Professional workevidence
Finance Agent (v2)
51%
Novel problemsevidence
ARC-AGI-2
no result
Long tasksno evidence yet
METR time horizon
no result

Third-party results only, at the model’s best standard setting. Hover a bar for who ran it.

What others sayrank on 8 leaderboards

Best attop three among current models

Not in the top three among current models on any category we track.

Where to use itprices checked 18 Nov 2026

Host prices from OpenRouter; going direct can differ.

Facts and sources

Released18 Sep 2026
Release entryStep 5 Preview
Scored asStep 5 Preview
Output speed–

StepFun releasesaround this one

This page as a chartDownload PNG