Step 3.7 Flash API Pricing
StepFun lists Step 3.7 Flash as a 198B-total / 11B-active sparse MoE model with 256K context, native image/video understanding, and tool calling. StepFun's USD price list charges $0.20/M input, $1.15/M output, and $0.04/M cached input. Below Step 5 Preview, StepFun's new flagship.
Run the numbers.
Calculator pre-loaded with the rates on StepFun's USD price list. Step 3.7 Flash is multimodal-capable; this page models the token pricing row.
Real-world presets.
Repo repair task
Analyst memo draft
Web research brief
Product support turn
This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| Step 3.7 Flash Current | $0.20 cache $0.04 | $1.15 | $0.155 agentic 92/8 | 256K | Cheaper StepFun multimodal agents |
| Step 5 Preview | $1.00 cache $0.05 | $2.70 | $0.419 +2.7× | 1M | StepFun's 1M-context agent flagship |
| Step 3.5 Flash | $0.10 cache $0.02 | $0.30 | $0.0556 −2.8× | 256K | Cheap StepFun text reasoning |
| MiMo-V2.6-Flash | $0.14 cache $0.0028 | $0.28 | $0.0477 −3.3× | 1M | Very low-cost 1M-context MiMo row |
| MiMo-V2.6-Pro | $0.435 cache $0.0036 | $0.87 | $0.144 −7% | 1M | MiMo Pro reasoning row |
| GLM-4.7 | $0.60 cache $0.11 | $2.20 | $0.358 +2.3× | 200K | Coding agent backend |
| DeepSeek V4 Pro | $1.32 cache $0.044 | $3.96 | $0.569 +3.7× | 1M | Low-cost reasoning workloads |
Frequently asked.
Practical questions about Step 3.7 Flash pricing, cache hits, context size, and StepFun multimodal reasoning workloads.
Q · 01 What is Step 3.7 Flash priced at? +
$0.20/M input on a cache miss, $0.04/M on a cache hit, and $1.15/M output for step-3.7-flash. The Chinese platform lists a separate CNY 1.35 / 0.27 / 8.1 card; the two are set independently, so this page reads the dollar list rather than converting the yuan one.Q · 02 How is the effective price calculated? +
92% input, 8% output, and 82% input cache hits. For Step 3.7 Flash, that lands at about $0.155/M blended tokens.Q · 03 What context window does Step 3.7 Flash support? +
256K context for step-3-7-flash. The model page also lists a sparse MoE architecture with 198B total / 11B active.Q · 04 Does this page model media/audio/image pricing? +
Q · 05 When was this price last checked? +
2026-09-29. Until then it converted the CNY card, which overstated the rates; see the price history and the corrections log.