Step 3.7 Flash

Open weights Active Multimodal MoE step-3-7-flash

198B MoE multimodal reasoning model (11B active) adding image and video input to Step 3.5 Flash.

Released
May 29, 2026
Context window
262K tokens
Max output
Not disclosed
Parameters
198B 11B active
Input price
$0.20 per 1M
Output price
$1.15 per 1M

Prices are first-party list prices in USD at the standard tier. Last verified Oct 30, 2026.

Pricing

Launch priceMay 29, 2026
$0.20 in / $1.15 out
Current price since May 29, 2026
$0.20 in / $1.15 out
Cached input $0.04 per 1M

Specifications

Developer
StepFun
Model type
Multimodal
Architecture
MoE
Parameters
198B (11B active)
Input
Text Image Video
Output
Text
Knowledge cutoff
Not disclosed
License
Apache License 2.0
API access
First-party API
Capabilities
Agentic Coding Reasoning Tool use Vision