Step 3.7 Flash
Last verified: June 17, 2026$0.200/1M input · $1.15/1M output · 262K context · StepFun
Count Tokens
—
Tokens
—
Input Cost
—
Output Cost
Estimate Monthly Cost
Monthly Cost Estimator
Quick:
< $0.0001/mo
Pick a preset above or enter custom usage
Alternatives to Step 3.7 Flash
Pricing Details
StepFunOpen-weight (Apache 2.0) multimodal MoE: 198B total / ~11B active, 196B language backbone + 1.8B ViT, 288 experts top-8. Cached input $0.04/1M. No surcharge for vision input or reasoning effort (low/medium/high dial). Native image + video, Python tool for hi-res image probing. AA Intelligence Index 43 (independent), ~382 tok/s, 0.94s TTFT. Vendor-reported: SWE-Bench Pro 56.3, Terminal-Bench 2.1 59.6, HR-Bench 4K 89.1. Advisor Mode claim: SWE-Bench Verified 76.3% at ~$0.19/task. HF stepfun-ai/Step-3.7-Flash. Also on OpenRouter and NVIDIA NIM (DeepInfra, Fireworks coming soon). Estimated tokens (different tokenizer). Max output unconfirmed.
Input / 1M tokens
$0.2
Output / 1M tokens
$1.15
Context Window
262K
Max Output
66K
Price History
Launched at current price on 2026-05-29. No price changes recorded.
Frequently Asked Questions
Common questions about Step 3.7 Flash pricing and usage