Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage & billing
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Usage & billing
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksNano RPCEarn crypto
  1. Models
  2. Text
  3. Stepfun Ai

Step 3.5 Flash

StepFun's most capable open-source reasoning model with visible reasoning traces. Built on a sparse Mixture-of-Experts architecture with 196B total parameters and only 11B active per token, it achieves frontier-level performance in math, logic, and agentic coding while reaching up to 350 tokens/sec. Supports 256K context. NOTE: This model runs via StepFun, which may log and train on your prompts.

  • Reasoning

Added Feb 2, 2026

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.10
Output
$0.30
Cache read
$0.05
Compare provider prices

Specifications

Context window
256K
Max output
256K
Parameters
196B / 11B
Total / active

Benchmarks

Sourced from LMArena.

Arena Score

1392.8

Overall Rank

#176 / 421

Votes

59,777

Confidence Interval

1389.3 - 1396.4

Category Scores

Coding

#162 / 416

15,074 votes

1449.0

Math

#155 / 404

3,403 votes

1405.7

Longer Query

#173 / 399

22,531 votes

1403.1

Creative Writing

#188 / 419

9,647 votes

1344.8

Instruction Following

#174 / 421

19,075 votes

1384.1

Hard Prompts

#181 / 421

37,189 votes

1412.2

Additional Categories
22

Japanese

#139 / 288

641 votes

1351.3

German

#140 / 320

1,069 votes

1393.1

Industry Mathematical

#141 / 401

3,116 votes

1419.9

Spanish

#145 / 303

2,005 votes

1403.8

Expert

#153 / 371

5,247 votes

1428.3

Korean

#157 / 289

1,073 votes

1339.7

French

#158 / 301

2,040 votes

1413.3

Chinese

#161 / 395

3,607 votes

1431.4

Industry Software And It Services

#165 / 421

22,592 votes

1437.0

Polish

#169 / 244

1,364 votes

1369.1

English

#175 / 421

25,049 votes

1408.8

Multi Turn

#175 / 419

10,329 votes

1396.1

Hard Prompts English

#176 / 419

16,111 votes

1423.1

Industry Business And Management And Financial Operations

#177 / 414

11,744 votes

1390.4

Exclude Ties

#178 / 421

43,530 votes

1371.2

Non English

#179 / 421

34,725 votes

1374.0

Industry Life And Physical And Social Science

#182 / 419

9,669 votes

1404.4

Industry Writing And Literature And Language

#183 / 420

14,620 votes

1365.4

Russian

#186 / 385

7,371 votes

1375.0

Industry Legal And Government

#187 / 393

4,612 votes

1394.8

Industry Entertainment And Sports And Media

#188 / 419

11,945 votes

1347.7

Industry Medicine And Healthcare

#190 / 390

4,184 votes

1398.1

Published 2026-11-09 · Matched as Step-3.5-Flash · CC BY 4.0

LMArena Dataset

Leaderboard results

BenchmarkScoreRank on NanoGPT
UGI Leaderboard

UGI Leaderboard data, Apache-2.0 · Updated Oct 16, 2026

UGI Score

40.8

#74 of 149

UGI Willingness

2.5

#98 of 149

UGI Writing

45.9

#66 of 149

Filtered and ranked to models on NanoGPT.

Providers

Auto routing is available for this model. Explicit provider selection is not available.

Auto
Auto
Recommended
$0.10$0.30$0.05N/AN/A
Batch API
ProviderInputOutputCached input
Auto
Autostepfun-ai/step-3.5-flash
$0.10/M$0.30/M$0.05/M

Batch rows for this model use normal routing and cost the same as standard API requests. Set provider in a row to choose a provider.

Reliability

Related text models

Compare Step 3.5 Flash with similar models from the same provider or model family.

Step 3.5 Flash 2603

stepfun-ai/step-3.5-flash-2603

Step 3.5 Flash 2603 is optimized for high-frequency agentic and coding workflows with improved token efficiency and faster reasoning. NOTE: This model runs via StepFun, which may log and train on your prompts.

Step 3.7 Flash Thinking

stepfun/step-3.7-flash:thinking

Step 3.7 Flash Thinking is StepFun's high-efficiency multimodal MoE model with visible reasoning enabled for deeper agentic coding, long-context reasoning, tool use, and native image/video understanding.

Step 5 Preview

stepfun/step-5-preview

Step 5 Preview is StepFun's 600B sparse MoE frontier model for production-scale agents, activating 27B parameters per token. It is built for software engineering, long-horizon tool use, research, professional knowledge work, and finance, with native text, image, and video understanding and a 1M-token context window.

OrcaCyber DeepSeek V4.1 Flash

orca/orcacyber-flash-1.0

An OrcaCyber model built on DeepSeek V4.1 Flash for authorized security research, coding, and tool use. Accepts text and images, with reasoning off by default. Supports a 1M-token context window, up to 393K output tokens, tool calling, and structured output. Offensive work is permitted only within an authorized security engagement.

OrcaCyber DeepSeek V4.1 Flash Thinking

orca/orcacyber-flash-1.0:thinking

An OrcaCyber model built on DeepSeek V4.1 Flash for authorized security research, coding, and tool use. Accepts text and images, with reasoning on by default. Supports a 1M-token context window, up to 393K output tokens, tool calling, and structured output. Offensive work is permitted only within an authorized security engagement.

OrcaCyber GLM 5.3 Flash

orca/orcacyber-glm-5.3-flash

An OrcaCyber model built on GLM 5.3 Flash for authorized security research, coding, and tool use. Accepts text and images. Reasoning is always enabled, with selectable effort. Supports a 1M-token context window, up to 131K output tokens, tool calling, and structured output. Offensive work is permitted only within an authorized security engagement.