Language modelOpen source

Step 3.5 Flash

StepFun

Released: February 2026
Data date: September 8, 2026
Successor: Step 3.7 Flash

Speed: 218.4 tokens/s
Output throughput
Price: $0.1 · $0.3
Input · Output / 1M tokens
Input: Text
Output: Text

Step 3.5 Flash combines roughly 197 billion total parameters with about 11 billion active and a 256K context using sliding-window attention. The fast open-weights model processes text; StepFun also evaluates a Parallel Thinking variant.

StepFun documents Step 3.5 as a text model and the direct predecessor of Step 3.7 Flash. The later release reports measurable gains on SWE-Bench Pro and Terminal-Bench 2.1.

Specifications and access

Specification	Value and source
Weights	Open weightsSource
Access	API, LocalSource
Parameters	196.8BSource
Active parameters	11BSource
Context window	256,000 tokensSource
Notes	Apache-2.0 model with 196.81B total parameters, about 11B active parameters, and a 256K context window.Source
License	Apache-2.0Source
Parameters	196.81 BSource
Active parameters	11 BSource
Architecture	45-layer sparse MoE Transformer, hidden size 4096, 288 routed experts plus one shared expert, top-8 routing, 3-way MTPSource
Input	TextSource
Output	TextSource
Reasoning mode	reasoning and agent scenarios recommend temperature 1.0, top_p 0.95; parallel thinking is documentedSource
Tool calling	YesSource
API model ID	step-3.5-flashSource
Throughput in tokens per second	100-300 typical, up to 350 single-stream coding (Throughput is a usage claim, not a benchmark score.)Source

Measurements without matching peer values

These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.

tau2-Bench (Show measurement, test conditions, and source)

Source value: 88.2
Score: 88.2
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

BrowseComp (Show measurement, test conditions, and source)

Source value: 51.6
Score: 51.6
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

BrowseComp (with Context Manager) (Show measurement, test conditions, and source)

Source value: 69
Score: 69
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

BrowseComp-ZH (Show measurement, test conditions, and source)

Source value: 66.9
Score: 66.9
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

BrowseComp-ZH (with Context Manager) (Show measurement, test conditions, and source)

Source value: 73.7
Score: 73.7
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

GAIA (no file) (Show measurement, test conditions, and source)

Source value: 84.5
Score: 84.5
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

xbench-DeepSearch (2025.05) (Show measurement, test conditions, and source)

Source value: 83.7
Score: 83.7
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

xbench-DeepSearch (2025.10) (Show measurement, test conditions, and source)

Source value: 56.3
Score: 56.3
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

ResearchRubrics (Show measurement, test conditions, and source)

Source value: 65.3
Score: 65.3
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

AIME 2025 (Show measurement, test conditions, and source)

Source value: 97.3
Score: 97.3
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

HMMT 2025 (Feb.) (Show measurement, test conditions, and source)

Source value: 98.4
Score: 98.4
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

HMMT 2025 (Nov.) (Show measurement, test conditions, and source)

Source value: 94
Score: 94
Metric: percent
Unit: %
Category: mixed
Direction: Higher is better
Source type: vendor-reported
Evaluator: StepFun Step 3.5 Flash repository/model README
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Official detailed table; benchmark-specific harness, context-manager and reproduced-score caveats are stated in the README.

Source StepFun Step 3.5 Flash repository/model README

Page 1 of 2

Predecessor and successor

Sources and data date

Type	Evidence and data date
Speed	StepFun. Output speed after the first streaming chunk, not total response time. Classified as deprecated by Artificial Analysis. Artificial Analysis maintains only the default 10K-input workload there. Retrieved September 8, 2026. Source: Artificial Analysis · Artificial Analysis (retrieved September 8, 2026)Speed measurement methodology
Price	Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs. Source: Pricing · API pricing
Input	Documented modalities. Retrieved September 8, 2026. Source: Model documentation · Model metadata (retrieved September 8, 2026) · StepFun Step 3.5 Flash repository · Editorial description reviewed October 2, 2026
Output	Documented modalities. Retrieved September 8, 2026. Source: Model documentation · Model metadata (retrieved September 8, 2026) · StepFun Step 3.5 Flash repository · Editorial description reviewed October 2, 2026
Research status	Research date September 8, 2026. 3 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Not found in the inspected sources	release-date Original wording The checked repository and release page do not expose a precise dated launch field.
Not found in the inspected sources	knowledge-cutoff Original wording The exact Step 3.5 Flash README does not publish a knowledge cutoff.
Additional source	StepFun Step 3.7 Flash release · Editorial description reviewed October 2, 2026
Additional source	OpenRouter: Popularity index (derived), CC BY 4.0
Additional source	Vercel AI Gateway: Popularity index (derived), CC BY 4.0
Additional source	Creative Commons Attribution 4.0

Step 3.5 Flash

Specifications and access

Measurements without matching peer values

Our tests

Pelican on a bicycle

Popularity

Published prices

Predecessor and successor

Sources and data date