What is StepFun best for?

It's the fast workhorse tier: long agent chains and large-file work where throughput and stability matter; 256K context fits large codebases. Pair it with a flagship model (DeepSeek, MiniMax M3) for peak single-answer intelligence.

SETUP GUIDE

How to run Claude Code & Codex with StepFun 阶跃星辰

StepFun's step-3.7 series is a high-speed model line built for agent workflows: 256K context, fast responses, and stable long tool-call chains, with official adaptations for Claude Code and other agent frameworks. Dual-engine ready (Anthropic-compatible + Responses API).

Engine support (probe-verified, 2026-07-10)

Engine	StepFun 阶跃星辰 support
Claude Code	✅ Supported (Anthropic-compatible endpoint)
Codex	✅ Supported (OpenAI Responses API)

Step 1 — Get a StepFun 阶跃星辰 API key

Open platform.stepfun.com → sign up / log in → “Interface keys” → create and copy.

Console: https://platform.stepfun.com/interface-key

Step 2 — Configure the Snail launcher (3 steps)

Download the launcher for Mac / Windows from justalittlesnail.ai;
Paste the API key on the StepFun 阶跃星辰 card (the launcher validates connectivity immediately);
Pick a model and click Launch Claude Code — a dedicated terminal window opens.

Recommended models

Main: step-3.7-flash (complex tasks) · Fast: step-3.7-flash (subtasks / summaries). After saving the key, click “Refresh” on the card to pull the live model list.

FAQ

What is StepFun best for?: It's the fast workhorse tier: long agent chains and large-file work where throughput and stability matter; 256K context fits large codebases. Pair it with a flagship model (DeepSeek, MiniMax M3) for peak single-answer intelligence.

Multi-window sessions, a different model in each window

Download the Snail launcher

Other guides: DeepSeek · Qwen 通义千问 · Kimi · GLM 智谱 · MiMo · MiniMax · 百度千帆 ERNIE · Anthropic Claude 官方
 All guides (EN) · 中文版 · Home