StepFun: Step 3.5 Flash

Text input Text output Free Option
Author's Description

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. It is a reasoning model that is incredibly speed efficient even at long contexts.

Key Specifications
Cost
$$$
Context
256K
Parameters
196B (Rumoured)
Released
Jan 29, 2026
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Temperature Stop Frequency Penalty Include Reasoning Reasoning Max Tokens Top P Tools
Features

This model supports the following features:

Tools Reasoning
Performance Summary

StepFun's Step 3.5 Flash, released on January 29, 2026, is a highly efficient open-source foundation model leveraging a sparse Mixture of Experts (MoE) architecture. It consistently ranks in the top tier for speed, performing in the 62nd percentile across nine benchmarks, and offers exceptionally competitive pricing, placing in the 81st percentile across eight benchmarks. The model demonstrates outstanding reliability with a 99% success rate, indicating minimal technical failures. Step 3.5 Flash exhibits perfect accuracy (100%) across a remarkable seven out of nine benchmarks, including Hallucinations, Instruction Following, General Knowledge, Email Classification, Reasoning, Ethics, and Mathematics. This highlights its robust capabilities in critical areas such as understanding uncertainty, complex instruction execution, and logical problem-solving. Notably, it achieves perfect accuracy in these categories while often being the most accurate model at its price point and speed. While its Coding performance is strong, achieving 94.9% and 91.9% accuracy, the duration for one Coding benchmark was significantly higher (33,404,092ms), suggesting potential variability or specific challenges in certain coding tasks. Overall, Step 3.5 Flash stands out for its exceptional accuracy, cost-efficiency, and speed, making it a powerful and reliable choice for a wide range of applications.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.1
Completion $0.3
Input Cache Read $0.02

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
StepFun
StepFun | stepfun/step-3.5-flash 256K $0.1 / 1M tokens $0.3 / 1M tokens
SiliconFlow
SiliconFlow | stepfun/step-3.5-flash 262K $0.1 / 1M tokens $0.3 / 1M tokens
DeepInfra
DeepInfra | stepfun/step-3.5-flash 262K $0.1 / 1M tokens $0.3 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration