StepFun

Step 3.7 Flash

Generally availableOpen weights262K contextReasoning

Summary

Step 3.7 Flash is an open-weight language model from StepFun, released on 29 May 2026. It accepts text, images and video and generates text, with a 262K-token context window and up to 230K output tokens per response. As of 19 Nov 2026, the lowest listed API price is $0.20 per million input tokens and $1.15 per million output tokens (Hugging Face Inference Providers), across 4 providers we track.

Newer StepFun flash model for faster agents, coding, and multimodal prompts

Last verified 19 Nov 2026. Every figure on this page links to its source.

Specifications

DeveloperStepFun
Release date29 May 2026 [models.dev]
StatusGenerally available [models.dev]
WeightsOpen weights [Hugging Face]
Licenseapache-2.0 [Hugging Face]
Parameters201B [Hugging Face]
Context window262K (262,144 tokens) [models.dev]
Max output230K tokens [models.dev]
Knowledge cutoff1 Mar 2026 [models.dev]
InputText, images and video [models.dev]
OutputText [models.dev]
Reasoning modeYes [models.dev]
Tool callingYes [models.dev]
Structured outputYes [models.dev]
Hugging Facestepfun-ai/Step-3.7-Flash [Hugging Face]

Step 3.7 Flash API pricing by provider

USD per million tokens. Sorted by input price.

ProviderInputOutputCached inputContextSource
Hugging Face Inference Providers
stepfun-ai/Step-3.7-Flash
$0.20$1.15–262Kmodels.dev, 19 Nov 2026
Novita AI
stepfun/step-3.7-flash
$0.20$1.15$0.04262Kmodels.dev, 19 Nov 2026
OpenRouter
stepfun/step-3.7-flash
$0.20$1.15$0.04256KOpenRouter, 19 Nov 2026
Vercel AI Gateway
stepfun/step-3.7-flash
$0.20$1.15$0.04256Kmodels.dev, 19 Nov 2026

Running Step 3.7 Flash locally

Estimated memory, from the parameter count and the architecture in config.json on Hugging Face. Includes about 0.8 GB of runtime overhead.

QuantizationWeightsTotal, 8K contextTotal, 32K context
Q4_K_M (4-bit)122 GB127 GB136 GB
Q8_0 (8-bit)214 GB219 GB228 GB
FP16 / BF16 (16-bit)403 GB407 GB417 GB

At 4-bit with an 8K conversation, Step 3.7 Flash needs about 127 GB, more than any single consumer graphics card; it needs a unified-memory machine, a multi-GPU setup or a split between graphics card and system RAM. Check your own hardware, or see Best computer for local AI on our sister site AIPCs.

Frequently asked questions

What is the context window of Step 3.7 Flash?

Step 3.7 Flash has a context window of 262,144 tokens (262K) and can generate up to 230,400 tokens in a single response, according to models.dev.

How much does the Step 3.7 Flash API cost?

Step 3.7 Flash is listed from $0.20 per million input tokens and $1.15 per million output tokens (Hugging Face Inference Providers). Prices last verified 19 Nov 2026.

Is Step 3.7 Flash open source?

Step 3.7 Flash is an open-weight model: its weights are published on Hugging Face as stepfun-ai/Step-3.7-Flash under the apache-2.0 license. Check the license terms before commercial use.

When was Step 3.7 Flash released?

StepFun released Step 3.7 Flash on 29 May 2026, according to models.dev.

Which providers offer Step 3.7 Flash?

We track Step 3.7 Flash on 4 providers: Hugging Face Inference Providers, Novita AI, OpenRouter, Vercel AI Gateway.

How much VRAM does Step 3.7 Flash need?

About 127 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 219 GB at 8-bit. That is more than any single consumer graphics card holds. These are estimates from the model's published parameter count and architecture.

Similar models from other labs

Comparable on output type, weights, release date, price and context window. Prices are the lowest listed, USD per million tokens.

ModelLabReleasedContextInput $/1MOutput $/1MWeights
Macaron V1 TallMindlab21 Jul 2026262K$0.25$1.43Open
Qwen3.6 35B-A3BAlibaba (Qwen)17 Apr 2026262K$0.10$0.90Open
Qwen3.5 35B-A3BAlibaba (Qwen)23 Feb 2026262K$0.14$1Open
Hy3 previewTencent (Hunyuan)20 Apr 2026262K$0.18$0.60Open
Hy3Tencent (Hunyuan)6 Jul 2026262K$0.13$0.53Open
Mistral Small 4Mistral AI16 Mar 2026262K$0.15$0.60Open

Change history

  • Context window changed from 256K to 262K
  • Max output changed from 256K to 230K
  • Input modalities changed from text, image to text, image, video
  • Input modalities changed from text, image, video to text, image
  • Context window changed from 262K to 256K
  • Context window changed from 256K to 262K
  • Input modalities changed from text, image to text, image, video
  • Input modalities changed from text, image, video to text, image
  • Max output changed from 230K to 256K
  • Context window changed from 262K to 256K
  • Context window changed from 256K to 262K
  • Input modalities changed from text, image to text, image, video
  • Max output changed from 256K to 230K
  • Input modalities changed from text, image, video to text, image
  • Context window changed from 262K to 256K
  • Context window changed from 256K to 262K
  • Input modalities changed from text, image to text, image, video
  • Input modalities changed from text, image, video to text, image
  • Context window changed from 262K to 256K
  • Max output changed from 230K to 256K
  • Max output changed from 256K to 230K
  • Input modalities changed from text, image to text, image, video
  • Context window changed from 256K to 262K
  • Input modalities changed from text, image, video to text, image
  • Max output changed from 230K to 256K

Sources