Model catalogue
StepFun
Step 3.7 Flash
stepfun/step-3.7-flash
Specification
Published by the vendor; the platform bills at the target's declared price.
Step 3.7 Flash
reasoning open weights toolsStepFun stepfun/step-3.7-flash
Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...
- Input $/1M
- $0.20
- Output $/1M
- $1.15
- Context
- 262K
- Max output
- 230K
- Released
- May 28, 2026
- Modalities
- textimagevideo
Intelligence index-
Coding index40
Agentic index-
Cached input reads $0.04 per 1M · weights on Hugging Face as stepfun-ai/Step-3.7-Flash
On this deployment
No routing target resolves to this model yet. Declare one and the router picks it up without retraining.
Not routable here yet
Add a target with metadata.model set to this id.
targets.yamlyaml
# targets.yaml - declare a target that runs Step 3.7 Flash
targets:
- id: step-3-7-flash
kind: llm
name: Step 3.7 Flash
metadata:
model: stepfun/step-3.7-flash # links the target to this catalogue entry
capabilities:
domains: [general]
supports_tools: true
context_window: 262144
cost:
usd_per_1k_tokens: 0.000675Price among StepFun models
Input list price per million tokens; Step 3.7 Flash is the most expensive of 2 listed.
- 1$0.10
- 2$0.20Step 3.7 Flash
More from StepFun
Other models by the same vendor in the catalogue.