Step 3.5 Flash

Name: Step 3.5 Flash
Author: StepFun

Released

2026-01-29

Last refreshed

2026-06-29

Status

Researched 135d ago

ProprietaryCommercial use: conditionalCodingAgentsLong context

Step 3.5 Flash is worth evaluating for coding, agents, and long context when its provider route and context window match the workload.

Use it for

Teams evaluating coding, agents, and long context
Workloads that can use a 256k context window
Buyers comparing 1 tracked provider route

Do not use it for

Vision or document-understanding workloads
Strict JSON or tool-calling flows

Specifications

Family: Step
Released: 2026-01-29
Context: 256k
Parameters: 196B (11B active)
Architecture: Mixture of Experts
Specialization: general
Openness: Proprietary
License: ProprietaryCommercial use: conditional
Weights: Not released
Code: Unknown
Training: Pretrained

Created by

StepFun

One of China's leading AI 'Six Tigers'.

Shanghai, China

Founded 2023

Website

Pricing

Output / 1M

$0.300

Input / 1M

$0.100

Cheapest of 1 route · OpenRouter

Providers(1)

OpenRouter

View 1 provider route

About

Step 3.5 Flash is StepFun's most capable open-source foundation model. Its sparse Mixture of Experts architecture activates only 11B of 196B total parameters per token, making it highly efficient at long contexts. Achieves 60.47% on SWE-Bench Verified and 56.2% on SWE-Pro, competing with frontier closed models on agentic coding benchmarks. Supports 256K token context.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ B

2 relevant benchmarks in the decision map.

Agents

Q/$ A

1 relevant benchmark in the decision map.

Long context

Included by capability and metadata signals in the decision map.

Provider price ladder

Compare API pricing across 1 providers for input and output tokens, batch, and cached reads when available.

Provider	Input / 1M	Output / 1M	Route
OpenRouter	$0.100	$0.300	Serverless

Capabilities

Reasoning

Benchmark peer barsfor Coding

SWE-bench VerifiedRank 44 of 81

Claude Fable 5

96.0

Claude Opus 5

96.0

Claude Mythos Preview

93.9

Claude Opus 4.8

88.6

Step 3.5 Flashcurrent

74.4

LiveCodeBenchRank 16 of 56

DeepSeek V4 Pro

93.5

Gemini 3.1 Pro Preview

91.7

DeepSeek V4 Flash

91.6

Qwen3.7-Max

91.6

Step 3.5 Flashcurrent

86.4

Benchmark scores(3)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.

Benchmark	Score	Version	Evaluation	Source
AIME 2025	97.3	From official StepFun blog post (accuracy%)Observed 2026-06-07	—	Source
LiveCodeBench	86.4	LiveCodeBench-V6 (accuracy%)Observed 2026-06-07	—	Source
SWE-bench Verified	74.4	From official StepFun blog post (resolved%)Observed 2026-06-07	—	Source