All tasks / Models
Step 5 Preview
StepFun's open-weights flagship MoE with 600B total and 27B active parameters, 1M-token context.
- StepFun
- Open weights
- Released 20 Sep 2026
Where it ranks
- AChat & assistantsModels · Local · 1,447 on LMArena Text Arena (overall, style control)
- BWritingModels · Local · 1,403 on LMArena Text Arena: Creative Writing (style control)
- CCodingModels · Local · 1,565 on LMArena WebDev / Code Arena (overall)
- AImage understandingModels · Local · 1,250 on LMArena Vision Arena (overall, style control)
- CWorkflow automationModels · Local · 51% on AutomationBench-AA (Artificial Analysis)
Facts
- Vendor
- StepFun
- Family
- Step
- Released
- 20 Sep 2026
- Weights
- Open: download and run it yourself
- License
- StepFun Model License 1.0
- Parameters
- 604B (27B active)
- GPU memory to run locally
- ~346 GB (4-bit ≈ 604 B × 0.57 GB + 1.5 GB ≈ 346 GB, estimate.)
- API price
- $1 in / $2.7 out per 1M tokens (OpenRouter listed price)
- Access
- Free (open weights)
- Last verified
- 19 Oct 2026
Price from openrouter.ai, 19 Oct 2026.
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| Artificial Analysis Intelligence Index | 44 | Step 5 PreviewRaw value from the page's embedded model data; the site displays it rounded to an integer. | 19 Oct 2026 | artificialanalysis.ai |
| Artificial Analysis MMMU-Pro (independent evaluation) | 76.4% | Step 5 PreviewRaw fraction 0.7636 from the model data embedded in https://artificialanalysis.ai/leaderboards/models, converted to percent; rank 45 of 124 entries with an MMMU-Pro result; no last-updated date shown. | 19 Oct 2026 | artificialanalysis.ai |
| AutomationBench-AA (Artificial Analysis) | 51% | Step 5 PreviewAutomationBench-AA Score (share of each task's objectives completed, a task with any guardrail violation scoring zero); raw value from the page's embedded model data (displayed rounded); rank 25 of 32; Tasks Completed 17.66%, Objectives Completed 78.52%; no last-updated date shown, so as_of = retrieval date. | 19 Oct 2026 | artificialanalysis.ai |
| LMArena Text Arena (overall, style control) | 1,447 | Step 5 PreviewFrom LMArena's official leaderboard-dataset (text_style_control, split latest, category overall, published 2026-10-15), which mirrors https://arena.ai/leaderboard/text. Rank 84; 7116 votes; 95% CI 1439.8-1454.8. | 15 Oct 2026 | huggingface.co |
| LMArena Text Arena: Creative Writing (style control) | 1,403 | Step 5 PreviewFrom LMArena's official leaderboard-dataset (text_style_control, split latest, category creative_writing, published 2026-10-15), which mirrors https://arena.ai/leaderboard/text/creative-writing. Rank 104 of 414; 1517 votes; 95% CI 1387.2-1419.2. | 15 Oct 2026 | huggingface.co |
| LMArena Vision Arena (overall, style control) | 1,250 | Step 5 PreviewFrom LMArena's official leaderboard-dataset (vision_style_control, split latest, category overall, published 2026-10-09), which mirrors https://arena.ai/leaderboard/vision. Rank 55 of 165; 1088 votes; 95% CI 1232.1-1268.4. | 9 Oct 2026 | huggingface.co |
| LMArena WebDev / Code Arena (overall) | 1,565 | Step 5 Preview (High)From LMArena's official leaderboard-dataset (webdev = Code Arena, split latest, category overall, published 2026-10-18), which mirrors https://arena.ai/leaderboard/webdev. Rank 37; 4603 votes; 95% CI 1555.6-1575.4. | 18 Oct 2026 | huggingface.co |
Where you can use it
- SOpenRouterRuntimes & gateways · Cloud
Earlier versions
- Step 3.5 Flash · —
Compare
Sources
- https://huggingface.co/stepfun-ai/Step-5-Preview
- https://openrouter.ai/api/v1/models
- https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset
- https://artificialanalysis.ai/leaderboards/models
- https://openrouter.ai/api/v1/models/stepfun/step-5-preview/endpoints
- https://artificialanalysis.ai/evaluations/automationbench-aa