StepFun Launches Step 5 Preview: 600B MoE with 27B Active, Open Weights on Oct 15
airesearch12 · x · 2026-09-20
StepFun officially introduced Step 5 Preview, its new flagship model for agentic work:
- Architecture: 600B total / 27B active MoE, 1M context with vision input
- Positioning: frontier-level performance in software engineering and professional knowledge work (notably finance), with sustained long-horizon execution
- Cost: substantially lower task cost at comparable intelligence; third-party commentary claims near GLM 5.3 / Kimi K3 quality for 65% cheaper
- Open weights: coming October 15
A direct challenge to leading closed models on agentic capability and price-performance from a Chinese lab.
Related event: StepFun Launches Step 5 Preview: 600B MoE Flagship, Open-Weights on Oct 15(12 posts)→
More from Models
- DiffusionGemma as Jev: single-pass parallel denoising decisions in ~0.2s on DGX Spark — bodonoghue85 · 2026-09-20
- Sonnet 5.2, Opus 5.2, Gemini 4 Pro reportedly in testing as release wave looms — kimmonismus · 2026-09-20
- Jev after days of real testing: weak Chinese, no reasoning, 30x faster than DeepSeek only in closed tasks — ZeYanjie · 2026-09-20
- LLM vs classical ML across 8 datasets: labeled data still favors SVM and XGBoost — Ok_Juggernaut2187 · 2026-09-20
- Researchers Use Claude Opus 5 to Hijack OpenAI Employee Accounts for Under $3,000 — FinanceYF5 · 2026-09-20
- NetEase Youdao open-sources Confucius4-R2T2 streaming ASR with 200-600ms latency — aigclink · 2026-09-20