StepFun drops open-weight 600B MoE: 27B active, 1M context, day-one vLLM and SGLang support
lmoroney · x · 2026-10-16
StepFun released Step 5 Preview on Hugging Face: a sparse mixture-of-experts model with 600B total parameters, 27B active per token, and up to 1M-token context, taking text+image input and text output.
- Positioned for agentic software engineering, long-horizon tool use, and professional knowledge work with a finance focus
- Day-one serving on vLLM and SGLang with OpenAI-compatible APIs; FP8 fits on 4×B300 or 8×H200, BF16 needs 8×B300 or 16×H200 across two nodes
- Sample commands wire in speculative decoding and tool-call parsers
- License is StepFun Model License 1.0, Apache-2.0-style with extra attribution and commercial terms
The author cautions that benchmark charts are vendor-reported numbers — test on your own tasks.
Related event: StepFun Open-Sources Step 5 Preview: 600B MoE with 1M-Token Context(21 posts)→
More from coding & agent
- Agent named 'Ilya' reportedly scores 1.1pt higher on benchmarks — daniel_mac8 · 2026-10-16
- Gemini CLI fix: substring matching inlined unrequested binary files into model context — priyanshu1976 · 2026-10-16
- Building AI agents on real healthcare claims data: the domain work that makes them work — alex_verem · 2026-10-16
- OpenAI stealth-testing mysterious internal model 'Aeon' in Codex desktop app and Dots — cedric_chee · 2026-10-16
- Vibe-coded iPhone Duo sniping bot built with Codex in 17 minutes, 41 tests passing — xiaohu · 2026-10-16
- Long-term multiplayer agent memory tool tops BEAM memory benchmark — CurieuxExplorer · 2026-10-16