Skip to content

StepFun · Chat

Step 3.5 Flash

StepFun 🇨🇳ChatStep familyOpen weights. One of its sources says the weights are publicly available.

Newer version: Step 3.7 Flash. We compared release dates and version names in our catalogue.

From Empiriolabs's listing:
StepFun text reasoning model for agents, coding, tool calling, and long-context analysis.
llminfo IQ
125.8
Maker's price
$0.15 blended / 1M
Max context
262K
Listings
24

Evidence and sources

International API storefront (platform.stepfun.ai) is Singapore-based; principal operating entity 上海阶跃星辰智能科技股份有限公司 (Shanghai).

🤗 stepfun-ai/Step-3.5-Flashllminfo IQ 125.8, ranked among chat models. From 118 benchmarks. Plenty of data behind it.How llminfo IQ is computedCoding IQ 106.2, 95% interval 96–116. From 8 benchmark families across 4 sources. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.How Coding IQ is computedGood value (72.1). Cheaper per task than its benchmark results predict. From 19 benchmarks with measured task costs. Plenty of data behind it.How Value for money worksReleased 1 Feb 2026Maker HQ: CN (source)
All rankings: #362 on OpenRouter this month #452 on OpenRouter this weekSources: StepFun API data, EmpirioLabs data, Models.dev data, Hugging FacePrimary data from StepFun API from EmpirioLabs: primary mode, description from Models.dev: weights availability from Hugging Face: release date

Advantages and trade-offs

Compared with 10 chat models, picked mainly by popularity and IQ

Advantages

Open weights available

The weights are public. 3 of the 8 peers that say either way have open weights.

We only make a claim when there's enough data on models of the same type. Updated 22 Nov, 06:07 UTC.Prices come from the model's maker where they sell it, otherwise their official partner, otherwise the market median. This comparison uses 7 maker's prices; 3 market median prices.

Input & output

as stated by the model's providers
Text
Text

Capabilities

Supported operations and provider-stated features

ReasoningFunction callingStructured outputPrompt cachingWeb search

Listings

 StepFunstep-3.5-flashSources: Models.dev data, StepFun API data256K$0.10cache read $0.02$0.30
 Flowdotstepfun/step-3.5-flash:free · freeSource: FlowDot data256K$0.00$0.00
 Eden AIInference by DeepInfradeepinfra/stepfun-ai/Step-3.5-FlashSource: Models.dev data262K$0.09cache read $0.02$0.30
 OpenRouterstepfun/step-3.5-flashSource: GenAI Prices dataNot published by the provider$0.09cache read $0.02$0.30
 Commandcodestepfun/Step-3.5-FlashSource: Models.dev data256K$0.09cache read $0.02$0.30
 Hugging FaceInference by DeepInfrastepfun-ai/Step-3.5-FlashSource: Hugging Face Inference Providers data262K$0.10$0.30
 Zenmuxstepfun/step-3.5-flashSources: Models.dev data, ZenMux data256K$0.10$0.30
 Empiriolabsstep-3-5-flashSources: EmpirioLabs data, LiteLLM data, Models.dev data262K$0.10cache read $0.02$0.30
 SiliconFlowstepfun-ai/Step-3.5-FlashSources: Models.dev data, SiliconFlow data262K$0.10$0.30
 Siliconflow Cnstepfun-ai/Step-3.5-FlashSource: Models.dev data262K$0.1046$0.3138
Show all 16 listings

Showing 10 of 16: the maker's own listing first, then the cheapest per 1M input tokens from each provider.

Also sold by 302ai, Aihubmix.

8 listings without a price are hidden.

Included in plans

MiniStepFunSubscription$6.99/moPlan page
PlusStepFunSubscription$9.99/moPlan page
ProStepFunSubscription$29/moPlan page
MaxStepFunSubscription$99/moPlan page

Benchmarks

Named effort levels run from lowest to highest. Select a result to see its source details.

Human preference

Arena ELO54 related benchmarks · 55 results
Arena ELO
Source: LMArena data · Nov 2026
Arena-Hard v2
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 5
Chinese
Source: LMArena data · Nov 2026
Creative Writing
Source: LMArena data · Nov 2026
English
Source: LMArena data · Nov 2026
Exclude Ties
Source: LMArena data · Nov 2026
Show 48 more resultsHide 48 results
Expert
Source: LMArena data · Nov 2026
French
Source: LMArena data · Nov 2026
German
Source: LMArena data · Nov 2026
Hard Prompts
Source: LMArena data · Nov 2026
Hard Prompts English
Source: LMArena data · Nov 2026
Industry Mathematical
Source: LMArena data · Nov 2026
Japanese
Source: LMArena data · Nov 2026
Korean
Source: LMArena data · Nov 2026
Longer Query
Source: LMArena data · Nov 2026
Multi Turn
Source: LMArena data · Nov 2026
Non English
Source: LMArena data · Nov 2026
Polish
Source: LMArena data · Nov 2026
Russian
Source: LMArena data · Nov 2026
Spanish
Source: LMArena data · Nov 2026
Style-Controlled
Source: LMArena data · Nov 2026
Text Factuality
Source: LMArena data · Nov 2026
Text Factuality - Chinese
Source: LMArena data · Nov 2026
Text Factuality - English
Source: LMArena data · Nov 2026
Text Factuality - Exclude Ties
Source: LMArena data · Nov 2026
Text Factuality - Expert
Source: LMArena data · Nov 2026
Text Factuality - French
Source: LMArena data · Nov 2026
Text Factuality - German
Source: LMArena data · Nov 2026
Text Factuality - Hard Prompts
Source: LMArena data · Nov 2026
Text Factuality - Japanese
Source: LMArena data · Nov 2026
Text Factuality - Korean
Source: LMArena data · Nov 2026
Text Factuality - Longer Query
Source: LMArena data · Nov 2026
Text Factuality - Multi Turn
Source: LMArena data · Nov 2026
Text Factuality - Non English
Source: LMArena data · Nov 2026
Text Factuality - Polish
Source: LMArena data · Nov 2026
Text Factuality - Russian
Source: LMArena data · Nov 2026
Text Factuality - Spanish
Source: LMArena data · Nov 2026

Math

AIME 20254 results
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 5 · PaCoRe
Step 3.5 technical report · Tables 12–13 · Python tool
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
HMMT Feb 254 related benchmarks · 8 results
HMMT Feb 25
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 5 · PaCoRe
Step 3.5 technical report · Tables 12–13 · Python tool
Hmmt Feb 2026
Source: Hugging Face Official Benchmarks data · Reported by the model's developer. The source relaying it did not verify it. claim
HMMT February 2026
Source: OpenEvals Leaderboard data · Reported by the model's developer. The source relaying it did not verify it. · Mar 2026
HMMT Nov 25
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 5 · PaCoRe
Step 3.5 technical report · Tables 12–13 · Python tool
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 5 · PaCoRe
Step 3.5 technical report · Tables 12–13 · Python tool
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated

Agentic

BrowseComp4 related benchmarks · 14 results
BrowseComp
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 17 · 200-item subset
Step 3.5 Flash official headline table · vanilla · with Context Manager
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 11 · 256K · no-tools baseline
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
Tool Gain
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 11 · 256K · tool gain
zh
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla · with Context Manager
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 11 · 256K · no-tools baseline
ZH · Tool Gain
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 11 · 256K · tool gain
GAIA2 results
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla · no file
Step 3.5 technical report · Table 11 · 256K · no-tools baseline
GAIA · Tool GainStep 3.5 technical report · Table 11 · 256K · tool gain · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Tau-benchSource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 16 · avg@8
Step 3.5 Flash official headline table · vanilla

Coding

Codeforces RatingStep 3.5 technical report · Table 15 · CF-Div2-Stepfun · pass@8 rating · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 5 · PaCoRe
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
RepoQAStep 3.5 technical report · Table 5 · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Reasoning

ARC-AGI-1Step 3.5 technical report · Tables 12–13 · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 5 · text-only
Step 3.5 technical report · Tables 12–13 · text-only
Source: OpenEvals Leaderboard data · Reported by the model's developer. The source relaying it did not verify it. · Mar 2026
Setting not stated

Instruction following

IFBenchStep 3.5 technical report · Table 5 · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Multi-ChallengeStep 3.5 technical report · Table 5 · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Multilingual

EKI EstonianSource: Keelemudelite mõõdupuu data docs

Science

GPQA Diamond6 results
Reported by SenseTime / OpenSenseNova · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
SenseNova 6.7 Flash-Lite official benchmark chart · chart setting
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 5
Step 3.5 technical report · Tables 12–13
Source: OpenEvals Leaderboard data · Reported by the model's developer. The source relaying it did not verify it. · Mar 2026
Setting not stated

Long context

LongBench v2Step 3.5 technical report · Table 5 · short + medium subsets · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
OpenAI MRCR v2 · 8-needleStep 3.5 technical report · Table 5 · 8-needle · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Knowledge

MMLU-Pro3 results
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 5
Source: Hugging Face Official Benchmarks data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated

Vision

Other

AndroidDaily HardStep 3.5 technical report · internal evaluations · Step 3.5 Flash + Step-GUI · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 15 · accuracy avg@8
Step 3.5 technical report · Table 5 · C++
FRAMES · OracleStep 3.5 technical report · Table 5 · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
ResearchRubricsStep 3.5 Flash official headline table · vanilla · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
StepFun Consulting & RecommendationsStep 3.5 technical report · internal evaluations · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · internal evaluations · Claude Code · avg@3 · prose precision
Step 3.5 technical report · internal evaluations · Claude Code · avg@3
TAU2BenchStep 3.5 Flash official headline table · vanilla · Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
WildClawBench6 related benchmarks · 6 results
WildClawBench
Source: Hugging Face Official Benchmarks data · Reported by the model's developer. The source relaying it did not verify it. claim
Multimodal (26 tasks)
Reported by WildClawBench / InternLM · via BenchBoard docs
OpenClaw · multimodal-input split · 26 tasks
Overall
Reported by WildClawBench / InternLM · via BenchBoard docs
OpenClaw · full 60-task suite
Pure-text (34 tasks)
Reported by WildClawBench / InternLM · via BenchBoard docs
OpenClaw · pure-text split · 34 tasks
Total Cost
Reported by WildClawBench / InternLM · via BenchBoard docs
OpenClaw · full 60-task suite
Total Elapsed
Reported by WildClawBench / InternLM · via BenchBoard docs
OpenClaw · full 60-task suite
xbench-DeepSearch4 related benchmarks · 6 results
xbench-DeepSearch · 2025.05
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 11 · 256K · no-tools baseline
2025.05 · Tool Gain
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 11 · 256K · tool gain
2025.10 · Tool Gain
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 technical report · Table 11 · 256K · tool gain
xbench-DeepSearch · 2025.10
Reported by StepFun · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Step 3.5 Flash official headline table · vanilla
Step 3.5 technical report · Table 11 · 256K · no-tools baseline

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats · Hugging Face official benchmark leaderboards; licence varies by benchmark · LMArena, CC BY 4.0 (licence) · BullshitBench by Peter Gostev, MIT (licence) · Keelemudelite mõõdupuu: Independent LLM leaderboard for Estonian. Eesti Keele Instituut. https://moodupuu.eki.ee, CC BY 4.0. Latest per-benchmark results selected with partial-run flags retained. (licence) · OpenEvals/leaderboard-data, MIT (licence) · MathArena (ETH SRI / INSAIT), MIT — cite dekoninck2026matharena (licence) · SnakeBench (Greg Kamradt / ARC Prize Foundation), MIT (licence)

Also known as: deepinfra/stepfun-ai/step-3.5-flash, empiriolabs/step-3-5-flash, empiriolabs/step-3-5-flash-2603, step-3-5-flash, step-3-5-flash-2603, step-3.5-flash-2603, stepfun-ai/step-3.5-flash, stepfun/step-3.5-flash, stepfun/step-3.5-flash:free

Like this one. Tap or hover any dotted value, label or icon to see where it comes from and how it is measured. open details when tapped or hovered. How this data works