Step 3.7 Flash
StepFun · Open weights · Released May 28, 2026 · stepfun/step-3.7-flash
Best $0.20 in / $1.15 out via StepFun · fp8 · 256K ctx at this price (headline 262K)
262K contextVisionToolsReasoningStructured outputPrompt cachingOpen weights
Available from 2 providersUSD / MTOK · PRICES 23 MINUTES AGO · SPEED 9 DAYS AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | UPTIME |
|---|---|---|---|---|---|---|---|---|
 StepFun1ST PARTY | $0.20 | $1.15 | $0.04 | 256K | fp8 | 3s | 229 tok/s | 100.0% |
 Novita | $0.20 | $1.15 | $0.04 | 262K | fp8 | 2.6s | 150 tok/s | 100.0% |
TTFT/TPS: median (P50) over trailing 504h · chat10k profile (~10K in / 200 out) · measured via OpenRouter routing · methodology v1.1 · measured 9 days ago
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Input price history · best listed price, observed
$0.20 · SEP 2, 2026$0.20 · TODAY
Pricing variants · USD / MTok in / out
STANDARD
$0.20 / $1.15
per MTok in / out
CACHE READ $0.04
Recent changesRSS ↗
Oct 2, 2026endpoint removed · deepinfra3 → 2
Sep 2, 2026price decreased · input0.2 → 0.16
Sep 2, 2026price decreased · output1.15 → 0.92
About
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...
Specifications
Context window262,144
Max output230,400
Modalities intext, image, video
Modalities outtext
LicenseOpen weights
ReleasedMay 28, 2026
TokenizerOther
WeightsHugging Face ↗
Open weights
safetensors
Related
Embed badge
[](https://modelindex.ai/models/stepfun/step-3.7-flash)Sources
OpenRouter API23 minutes ago
AGGREGATORHugging Face23 minutes ago
COMMUNITYModelIndex benchmarks13 days ago
MEASUREDVerified 23 minutes ago · list prices, not negotiated rates.