DataLearner logo
ST

Step 3.7 Flash

Reasoning modelTool use

Step 3.7 Flash

Release date: 2026-05-29Updated: 2026-08-23Views: 1,311
Live demoGitHubHugging FaceCompare
Parameters
198B
Context length
256K
Multilingual
Supported
Reasoning ability
3/5

Step 3.7 Flash is a reasoning model from StepFunAI, released on 2026-05-29. It accepts text, image, and video input and returns text output. The cataloged parameter count is 198B, with 11B active parameters per inference. The recorded context window is 256K. Cataloged capabilities include Reasoning model and Multilingual. The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Step 3.7 Flash

Model basics

Reasoning traces
Supported
Thinking modes
Standard Mode (Default)Thinking Level · Deep
Context length
256K tokens
Max output length
No data
Model type
Reasoning model
Modality (in / out)
Text, Image, Video → Text
Release date
2026-05-29
Model file size
403 GB (BF16 Safetensors)
MoE architecture
Yes
Total params / Active params
198B / 11B
Knowledge cutoff
No data
Step 3.7 Flash

Open source & experience

Code license
Apache 2.0
Weights license
Apache 2.0- Commercial use permitted
GitHub repo
github.com/stepfun-ai/Step-3.7-Flash
Hugging Face
huggingface.co/stepfun-ai/Step-3.7-Flash
Live demo
studio.stepfun.com
Step 3.7 Flash

Official resources

Paper
Paper
DataLearnerAI blog
N/A
Step 3.7 Flash

API details

API speed
3/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Learn about pricing modes
Standard
TypeConditionInputOutput
Text-¥1.35/ 1M tokens¥8.10/ 1M tokens
Image-¥1.35/ 1M tokens—
Video-¥1.35/ 1M tokens—
Cache PricingPrompt Cache
TypeTTLWriteRead
Text-—¥0.270/ 1M tokens
Image-—¥0.270/ 1M tokens
Video-—¥0.270/ 1M tokens

“—” means the modality is not billed in that direction, or the vendor has not published a price for it.

Step 3.7 Flash

Benchmark Results

Step 3.7 Flash currently shows benchmark results led by τ²-Bench - Telecom (4 / 264, score 98.50), HLE (69 / 563, score 47.20), Terminal Bench Hard (70 / 244, score 35.60). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

3 evaluations
Benchmark / mode
Score
Rank/total
HLE
Thinking Mode
21.40
277 / 563
HLE
Thinking ModeTools
47.20
69 / 563
CritPt
Thinking Mode
2.30
126 / 200

General Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Thinking Mode
80.90
225 / 462

Multimodal Understanding

2 evaluations
Benchmark / mode
Score
Rank/total
SimpleVQA
Thinking ModeTools
79.20
1 / 3
MMMU-Pro
Thinking Mode
75.30
86 / 227

Coding and Software Engineer

2 evaluations
Benchmark / mode
Score
Rank/total
SWE-Bench Pro - Public
Thinking ModeTools
56.30
28 / 62
SciCode
Thinking Mode
43.90
93 / 130

Agent Level Benchmark

3 evaluations
Benchmark / mode
Score
Rank/total
τ²-Bench - Telecom
Thinking ModeTools
98.50
4 / 264
Terminal Bench Hard
Thinking ModeTools
35.60
70 / 244
τ³-Banking
Thinking ModeTools
12
123 / 164

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
Thinking Mode
67.30
90 / 282

AI Agent - Information Search

1 evaluations
Benchmark / mode
Score
Rank/total
BrowseComp
Thinking ModeTools
75.82
27 / 57

Text Embedding

3 evaluations
Benchmark / mode
Score
Rank/total
Context Arena
Low
33.97
108 / 126
Context Arena
Medium
37.65
101 / 126
Context Arena
High
35.33
104 / 126

AI Agent - Tool Usage

1 evaluations
Benchmark / mode
Score
Rank/total
Terminal-Bench 2.1
Thinking ModeTools
59.50
112 / 192

Productivity Knowledge

1 evaluations
Benchmark / mode
Score
Rank/total
Harvey Lab-AA
Thinking ModeTools
72.73
32 / 43
View benchmark analysisCompare with other models

Compare with other models

Step 3.7 Flash

Publisher

Step 3.7 Flash

Model Overview

Step 3.7 Flash is a reasoning model from StepFunAI, released on 2026-05-29.

It accepts text, image, and video input and produces text output. Its cataloged capabilities include Reasoning model and Multilingual. The cataloged parameter count is 198B, with 11B active parameters per inference. The recorded context window is 256K.

The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. The page records 7 API pricing rules from StepFunAI; current provider pricing and conditions should be checked before deployment. The evaluation section contains 5 cataloged benchmark results with their recorded modes and scores. The page links 4 release, model-card, repository, or provider references for checking the underlying claims. Specifications, availability, and prices can change; undisclosed values are intentionally left unstated.

Step 3.7 Flash

FAQ

What is Step 3.7 Flash?

Step 3.7 Flash is a reasoning model from StepFunAI, released on 2026-05-29. It accepts text, image, and video input and returns text output. The cataloged parameter count is 198B, with 11B active parameters per inference. The recorded context window is 256K. Cataloged capabilities include Reasoning model and Multilingual. The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

What input and output modalities does Step 3.7 Flash support?

The current model record lists text, image, and video as input and text as output.

What are the main recorded specifications for Step 3.7 Flash?

The cataloged parameter count is 198B, with 11B active parameters per inference. The recorded context window is 256K. Fields without a source-backed value remain undisclosed.

Does Step 3.7 Flash have API pricing?

The page records 7 API pricing rules from StepFunAI; current provider pricing and conditions should be checked before deployment.

Are benchmark results available for Step 3.7 Flash?

The evaluation section contains 5 cataloged benchmark results with their recorded modes and scores. Compare only results that use the same benchmark version and evaluation mode.

Is Step 3.7 Flash open source?

The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Review the linked license text before commercial or derivative use.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code