StepFun: Step 3.5 Flash

StepFun: Step 3.5 Flash

stepfun · Released Jan 29, 2026 Professional
Intelligence #14 / 590
82.0 Our Score
Speed #43 / 279
184.4 tokens / sec
Input #195 / 592
$0.100 per 1M tokens
Output #213 / 592
$0.300 per 1M tokens
Context #112 / 592
262,144 tokens

Analysis Summary

Step 3.5 Flash from StepFun is a well-rounded model with an intelligence index of 26 and a standout agentic index of 60, placing it ahead of many more expensive models on multi-step task execution. The 262K context window, tool use, and function calling make it capable across a range of business workflows. GPQA of 0.826 and a tau2 score of 0.874 indicate reliable instruction following and task completion.

For businesses, the combination of strong agentic performance and very low pricing ($0.10 input, $0.30 output) makes it attractive for high-volume automation, structured data extraction, and agent pipelines where cost efficiency matters. Long-context reasoning (LCR 0.543) is adequate for mid-length documents. The main limitation is that general reasoning depth sits below the top tier.

This is a strong value pick for agentic and tool-use workflows where budget is a constraint and top-tier reasoning is not required.

Assessed June 30, 2026

Editorial notes

StepFun Step 3.5 Flash delivers strong agentic performance (60 agentic index) and good reasoning at $0.10 input and $0.30 output, with a 262K context and tool use support.

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

Performance Profile

Intelligence4.5Technical7.3Value8Content6.9
Intelligence 4.5/10
Technical 7.3/10
Content 6.9/10
Value 8/10

How StepFun: Step 3.5 Flash compares

StepFun: Step 3.5 Flash ranks #98 of 385 AI models we track for overall intelligence, #61 of 293 for agentic tasks. Its 262K-token context window is larger than 81% of the models we list. At $0.10 per million input tokens it is cheaper than 67% of comparable models.

About StepFun: Step 3.5 Flash

Step 3.5 Flash is StepFun's most capable open-source foundation model. Built on a sparse Mixture of Experts (MoE) architecture, it selectively activates only 11B of its 196B parameters per token..

Capabilities

Tool Use Function Calling

Performance Indices

Source: Artificial Analysis

26 Intelligence Index
60 Agentic Index

Benchmark Scores

Intelligence

GPQA Diamond 82.6% Graduate-level scientific reasoning
HLE 22.6% Humanity's Last Exam
SciCode 38.5% Scientific computing

Technical

TerminalBench Hard 32.6% Agentic terminal tasks
τ²-Bench 87.4% Conversational agent benchmark

Content

IFBench 66.5% Instruction following
LCR 54.3% Long-context reasoning

Benchmark data from Artificial Analysis and Hugging Face

How does StepFun: Step 3.5 Flash stack up?

Compare side-by-side with other professional models.

Compare Models

Model Information

OpenRouter ID stepfun/step-3.5-flash
Providerstepfun
Release Date January 29, 2026
Context Length262,144 tokens
Max Completion65,536 tokens
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.10 $0.000100
Output $0.30 $0.000300

Live Performance

Live endpoint metrics, refreshed every 30 minutes.

100%
Avg Uptime
2,435ms
Best Latency (TTFT)
13 tok/s
Best Throughput
1/1
Active Endpoints
Available via: SiliconFlow

Leaderboard Categories

Frequently asked questions about StepFun: Step 3.5 Flash

How much does StepFun: Step 3.5 Flash cost?

StepFun: Step 3.5 Flash costs $0.10 per million input tokens and $0.30 per million output tokens.

What is the context window of StepFun: Step 3.5 Flash?

StepFun: Step 3.5 Flash has a context window of 262,144 tokens (262K).

What can StepFun: Step 3.5 Flash do?

StepFun: Step 3.5 Flash supports tool use and function calling.

Who created StepFun: Step 3.5 Flash?

StepFun: Step 3.5 Flash is developed by StepFun and was released on January 29, 2026.