One agent session
1M input tokens and 100K output tokens, no caching.
- Input1M x $0.10 = $0.10
- Output0.1M x $0.30 = $0.03
$0.13per session
StepFun - Closed weights
Step 3.5 Flash 2603 pricing and specs
Step 3.5 Flash 2603 is a closed-weights model from StepFun that costs $0.10 per 1M input tokens and $0.30 per 1M output tokens on the first-party API, with a 256K-token context window.
Last verified from platform.stepfun.ai. All sources
- Input / 1M
- $0.10
- Output / 1M
- $0.30
- Context
- 256K tokens
- Weights
- Closed
Pricing per 1M tokens
| Item | Price / 1M tokens |
|---|---|
| Input | $0.10 |
| Output | $0.30 |
| Cache read | $0.02 |
Note: StepFun international platform, priced in USD (same price as step-3.5-flash). Agent-optimized tune of Step 3.5 Flash with better token efficiency; reasoning_effort accepts low and high only. Context listed as 256K tokens (recorded as 256,000). Marked closed because no official weights repo for the 2603 variant was found (only the base Step-3.5-Flash repo is public). max output and release date blank: no primary page states them. Not listed in the Vercel AI Gateway model list on 2026-10-17.
Specs
- Context window
- 256,000 tokens
- Provider
- StepFun
- API model id
step-3.5-flash-2603- Weights
- Closed weights
What Step 3.5 Flash 2603 costs in practice
Worked from the standard list rates above, with the math shown so you can swap in your own token counts.
A chat app
10,000 requests a day at 2K input and 500 output tokens each.
- Input20M x $0.10 = $2.00
- Output5M x $0.30 = $1.50
$3.50per day
$105.00 per 30-day month
The same session at a 90% cache hit
90% of the input is read from the prompt cache. The one-time cache write premium is not included.
- Cached input0.9M x $0.02 = $0.018
- Uncached input0.1M x $0.10 = $0.01
- Output0.1M x $0.30 = $0.03
$0.058per session
Compare with
| Model | Input / 1M | Output / 1M |
|---|---|---|
| Step 3.5 Flash 2603 (this page) | $0.10 | $0.30 |
| Step 3.5 FlashStepFun | $0.10 | $0.30 |
| Hy-MT2-PlusTencent | $0.074 | $0.295 |
| Hy-MT2-ProTencent | $0.074 | $0.295 |
| MiMo-V2.6-FlashXiaomi MiMo | $0.14 | $0.28 |
| Step 3.7 FlashStepFun | $0.20 | $1.15 |
| Step-1o Turbo VisionStepFun | $0.36 | $1.15 |
| Step 5 PreviewStepFun | $1 | $2.70 |
Related reading
- Step 5 Preview: Pricing, Benchmarks and the October 15 Open WeightsStep 5 Preview is StepFun's 600B MoE with 27B active and a 1M window. What it scores, what a session costs, and what to check before the October 15 weights.
- Build Your First Agent with Vercel eve: A Step-by-Step TutorialBuild your first AI agent with Vercel eve: scaffold a project, add a typed tool with defineTool, run it locally, call the session API and deploy to Vercel.
- Agent Architecture: Building Multi-Step AI Workflows That Survive ProductionAgent architecture for multi-step Claude workflows: loop patterns, state management, error recovery, and the production gotchas that break agents at scale.
- The Agent Reliability Cliff: Why Your 10-Step Chain Only Succeeds 20% of the TimeThe agent reliability cliff: 85% per-step success compounds to about 20% over 10 steps. Why long chains collapse, and six patterns that fight the decay.
FAQ
How much does Step 3.5 Flash 2603 cost per million tokens?
Step 3.5 Flash 2603 costs $0.10 per 1M input tokens and $0.30 per 1M output tokens on StepFun's first-party API, per platform.stepfun.ai on October 17, 2026. Per 1M tokens, cache reads cost $0.02.
What is Step 3.5 Flash 2603's context window?
Step 3.5 Flash 2603 has a 256,000-token context window.
Are Step 3.5 Flash 2603's weights open?
No. Step 3.5 Flash 2603 is a closed-weights model. You call it through StepFun's API as step-3.5-flash-2603.
Sources
Last verified from platform.stepfun.ai.
Newsletter
Get model price changes as they ship
One email per week. New model releases, price cuts and what they mean for your bill.