Skip to content

StepFun Step 3.7 Flash + OpenCode: Cheapest Reasoning Setup

Set up OpenCode with StepFun Step 3.7 Flash for a very cheap, multimodal reasoning coding agent. Provider config, model IDs, low pricing, and the common fixes.

MGMCSA Guru Team July 20, 2026 3 min read
OpenCode configured with StepFun Step 3.7 Flash on Windows

StepFun’s Step 3.7 Flash is one of the cheapest reasoning-capable models you can put behind a coding agent — a multimodal mixture-of-experts design that activates only a small slice of its parameters per token, keeping it fast and inexpensive. Paired with free, open OpenCode, it makes a low-cost agent that can also reason over images if you need that.

This is the OpenCode + StepFun setup. For OpenCode’s general provider model, see OpenCode custom providers.

Step 1: Install OpenCode and get a StepFun key

npm install -g opencode-ai

Sign up on the StepFun platform and create an API key (its quickstart covers key creation). On Windows, WSL gives the cleaner shell — see install OpenCode on Windows with WSL.

Step 2: Add the StepFun provider

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "stepfun": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "StepFun",
      "options": {
        "baseURL": "https://api.stepfun.com/v1",
        "apiKey": "{env:STEPFUN_API_KEY}"
      },
      "models": {
        "step-3.7-flash": { "name": "Step 3.7 Flash" },
        "step-3.5-flash": { "name": "Step 3.5 Flash" }
      }
    }
  },
  "model": "stepfun/step-3.7-flash"
}

Export the key:

export STEPFUN_API_KEY="your-stepfun-key"

Confirm the exact base URL and model IDs on StepFun’s docs.

Step 3: Run it

opencode

The default is stepfun/step-3.7-flash. For the lowest cost on simpler tasks, switch to step-3.5-flash.

StepFun models (verify rates on StepFun)

step-3.7-flash Tunable reasoning; multimodal; very cheap
step-3.5-flash Even cheaper; great for simple tasks

Pricing

StepFun’s pricing is at or below the floor most mid-tier providers held — Step 3.7 Flash in the region of $0.20 input / $1.15 output per million, and Step 3.5 Flash even lower (around $0.10 / $0.30), illustrative. Multimodal images bill at a fixed token count each. Confirm current rates on StepFun’s pricing page.

Troubleshooting

  • Provider not found — the provider key (stepfun) must match the model prefix.
  • Auth errorsSTEPFUN_API_KEY not set in the launching shell.
  • Model rejected — confirm the Step model ID on StepFun.
  • Format errors — confirm you’re using StepFun’s OpenAI-compatible endpoint.

StepFun + OpenCode checklist

  • OpenCode installed
  • StepFun API key created and exported
  • Provider block with StepFun's OpenAI base URL
  • Step 3.7/3.5 Flash IDs listed; default set
  • Launched with opencode

Wrapping up

StepFun Step 3.7 Flash with OpenCode is about the cheapest reasoning-capable coding agent you can assemble: a free, open tool on a fast, multimodal MoE model at floor-level pricing. Add the provider block, set step-3.7-flash as default (drop to 3.5 Flash for simple work), and launch. Verify the endpoint and rates on StepFun’s docs.

For the broader budget picture, see cheapest AI coding API in 2026.

Frequently asked questions

What is StepFun Step 3.7 Flash?

It's StepFun's high-efficiency multimodal mixture-of-experts model — a reasoning-capable model that activates only a fraction of its parameters per token, so it's fast and very cheap. It pairs a language backbone with a vision encoder for image and video understanding.

How cheap is it?

Very — StepFun's rates are at or below the cost floor most mid-tier providers held, with Step 3.5 Flash even lower than 3.7. Confirm the current per-million input/output rates on StepFun's pricing page.

How do I add StepFun to OpenCode?

Add a custom provider in opencode.json using @ai-sdk/openai-compatible with StepFun's OpenAI-compatible base URL and your API key, then list the Step model ID and set it as default.

Is it good for coding?

It's a capable, cheap reasoning model that works for everyday coding, and its multimodal ability is a bonus if you feed it images. For the hardest tasks, a dedicated coding flagship may edge it, but for the price it's strong.

Does it run on Windows?

Yes, and cleanly in WSL. The OpenCode config is identical either way; WSL just gives the smoother terminal.

Sources & further reading

Official vendor documentation referenced while writing this guide.

MG

MCSA Guru Team

IT & Systems Administration

We are working IT pros and system administrators who spend our days in Windows Server, Microsoft 365, and the wider Microsoft stack. MCSA Guru is where we write down the fixes and walkthroughs we wish we had found the first time.

MCSA Guru provides independent, educational IT guidance. Microsoft, Windows, Windows Server, Microsoft 365, Exchange, and Microsoft Teams are trademarks of Microsoft Corporation; Docker is a trademark of Docker, Inc. MCSA Guru is not affiliated with or endorsed by Microsoft or Docker. Always test changes in a safe environment before applying them in production.

Related guides

Fixing something right now?

Jump straight into the guide library or search for the exact error or task you are dealing with.