Step 3.7 Flash
Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters per token. The model supports a 256K context window and exposes selectable reasoning levels (high/medium/low), letting callers trade off speed, cost, and depth of reasoning.
Designed for coding, agentic workflows, structured outputs, and long-context productivity tasks.
- Organization
- StepFun (China)
- Family
- —
- Providers
- 21
- Context
- 256,000
- Output limit
- 256,000
- Knowledge
- 2026-03-01
- Release
- 2026-05-29
- Updated
- —
- Weights
- Open
- Input
- text, image/jpeg, image/png, image/gif, image/webp, video/mp4, video/quicktime, video/x-matroska
- Output
- text
- Capabilities
- —
Providers
Compare every cataloged offer for this model, including its provider, route, token limits, price, and capabilities.
| Provider | Model ID | Context | Output | Input / output · 1M | Reasoning | Tools | Structured | Details |
|---|---|---|---|---|---|---|---|---|
| AIHubMix | step-3.7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
| CommandCode | stepfun/Step-3.7-Flash | 256,000 | 256,000 | — / — | — | — | — | View |
| Deep Infra | stepfun-ai/Step-3.7-Flash | 262,144 | 256,000 | — / — | — | — | — | View |
| DevPass (LLM Gateway) | step-3.7-flash | 262,144 | 256,000 | — / — | — | — | — | View |
| Eden AI | flexai/Step-3.7-Flash | 262,144 | 256,000 | — / — | — | — | — | View |
| Eden AI | deepinfra/stepfun-ai/Step-3.7-Flash | 262,144 | 256,000 | — / — | — | — | — | View |
| EmpirioLabs AI | step-3-7-flash | 256,000 | 131,072 | — / — | — | — | — | View |
| Hugging Face | stepfun-ai/Step-3.7-Flash | 262,144 | 256,000 | — / — | — | — | — | View |
| Kenari | step-3-7-flash:free | 256,000 | 256,000 | — / — | — | — | — | View |
| Kenari | step-3-7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
| Kilo Gateway | stepfun/step-3.7-flash | 256,000 | 230,400 | — / — | — | — | — | View |
| LLM Gateway | novita/step-3.7-flash | 262,144 | 256,000 | — / — | — | — | — | View |
| NanoGPT | stepfun/step-3.7-flash:thinking | 262,144 | 256,000 | — / — | — | — | — | View |
| Novita AI | stepfun/step-3.7-flash | 262,144 | 256,000 | — / — | — | — | — | View |
| Novitavia OpenRouter | stepfun/step-3.7-flash fp8 | 262,144 | 256,000 | $0.20 / $1.15 | Yes | Yes | Yes | View |
| StepFunvia OpenRouter | stepfun/step-3.7-flash fp8 | 256,000 | 230,400 | $0.20 / $1.15 | Yes | Yes | Yes | View |
| Requesty | step-3.7-flash | 262,144 | 256,000 | — / — | — | — | — | View |
| StepFun (China) | step-3.7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
| StepFun (Global) | step-3.7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
| StepFun Step Plan (China) | step-3.7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
| StepFun Step Plan (Global) | step-3.7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
| UnoRouter | step-3.7-flash:free | 256,000 | 256,000 | — / — | — | — | — | View |
| Vercel AI Gateway | stepfun/step-3.7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
| ZenMux | stepfun/step-3.7-flash | 256,000 | 256,000 | — / — | — | — | — | View |
Prices and limits apply to each offer. A dash means the value is not provided. OpenRouter offers show the hosting provider and routing channel separately.
Pricing
Compare token rates and cache charges across available routes and hosting configurations.
Lowest listed input and output prices across the offers above; they may belong to different providers. These are base rates, before conditional pricing, discounts or additional fees.
stepfun/step-3.7-flashObserved - Input
- $0.20
- Output
- $1.15
- Cache read
- $0.04
Hosting prices through OpenRouter All endpoints and cache rates
| Endpoint | Input / 1M | Output / 1M | Cache read / 1M | Cache write / 1M |
|---|---|---|---|---|
| Novitanovita/fp8 | $0.20 | $1.15 | $0.04 | — |
| StepFunstepfun/fp8 | $0.20 | $1.15 | $0.04 | — |
Each row is a hosting configuration. Open an offer in Providers for its full pricing conditions and additional fees. Endpoint observations may differ from the routing catalog's base price.
Performance
Review reported latency and output-throughput percentiles from the latest observed rolling window.
Time to first token · milliseconds
All latency percentiles 2 endpoints
| Endpoint | P50 | P75 | P90 | P99 |
|---|---|---|---|---|
| Novitanovita/fp8 | 3,858 ms | 5,800 ms | 7,733.4 ms | 11,501.92 ms |
| StepFunstepfun/fp8 | 3,470 ms | 4,937 ms | 7,408.2 ms | 13,053.8 ms |
Output throughput · tokens per second
All throughput percentiles 2 endpoints
| Endpoint | P50 | P75 | P90 | P99 |
|---|---|---|---|---|
| Novitanovita/fp8 | 48 t/s | 73.5 t/s | 95 t/s | 142.08 t/s |
| StepFunstepfun/fp8 | 79 t/s | 92 t/s | 107 t/s | 166.74 t/s |
Percentiles describe the observed request distribution. These are snapshots of a rolling window, not a historical time series.
Uptime
Compare successful request rates across the latest 5-minute, 30-minute, and 24-hour observations.
Reported by OpenRouter at the time of collection; rate-limited requests are excluded. A dash means no measurement was reported. This is not a live availability check.
Benchmarks
Review attributed evaluation results on each source's original scale.
stepfun/step-3.7-flashObserved | Design Arena | Category | Elo | Win rate | Rank |
|---|---|---|---|---|
| models | 3d | 1,134 | 40.6% | 71 |
| models | asciiart | 1,158 | 45% | 38 |
| models | codecategories | 1,185 | 42.6% | 59 |
| models | dataviz | 1,174 | 42.9% | 65 |
| models | gamedev | 1,166 | 38.9% | 66 |
| models | svg | 1,080 | 37.1% | 64 |
| models | uicomponent | 1,184 | 42.2% | 59 |
| models | website | 1,199 | 44.2% | 61 |
Sources: Artificial Analysis and Design Arena, via OpenRouter. Scores retain their original scales and are not OpenAI Suite ratings.
Detailed evaluations
| Evaluation | Configuration | Source | Results | Details |
|---|---|---|---|---|
| Composite indices | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Artificial Analysis | Intelligence 19.5 · Coding 39.6 · Agentic — | Details |
| models · uicomponent | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,176 Elo | Details |
| models · dataviz | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,173 Elo | Details |
| models · website | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,195 Elo | Details |
| models · 3d | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,130 Elo | Details |
| models · asciiart | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,153 Elo | Details |
| models · codecategories | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,182 Elo | Details |
| models · svg | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,083 Elo | Details |
| models · gamedev | Step 3.7 Flashstepfun/step-3.7-flash-20260528 | Design Arena | 1,159 Elo | Details |
Sources: Artificial Analysis, Design Arena and OpenRouter. Evaluation configurations and scales differ; results are not interchangeable. Updated Oct 24, 2026.
Apps & session costs
Compare observed median session costs for applications using this model, grouped by conversation length.
| Application | Model route | Turns | Median cost / session | Details |
|---|---|---|---|---|
| Hermes Agent | stepfun/step-3.7-flash-20260528 | 2–9 turns | $0.007758 | Details |
| Hermes Agent | stepfun/step-3.7-flash-20260528 | 10–49 turns | $0.054871 | Details |
Source: OpenRouter, as of Oct 20, 2026. Published under CC BY 4.0. These are observed costs per session, not token prices or a forecast for your workload. Apps are compared separately.
Activity
Daily prompt and completion tokens reported for this model's routes in OpenRouter's top 50. Missing days have no published value.
227,222,837,235 tokens
266,482,278,634 tokens
324,874,967,863 tokens
282,473,777,078 tokens
202,112,542,879 tokens
177,929,114,876 tokens
174,724,791,518 tokens
195,693,794,217 tokens
188,156,210,411 tokens
198,991,386,479 tokens
183,760,329,848 tokens
177,898,765,824 tokens
164,102,567,452 tokens
158,346,687,762 tokens
171,499,623,423 tokens
170,094,087,852 tokens
130,992,558,780 tokens
142,809,650,019 tokens
125,957,921,257 tokens
111,007,985,109 tokens
113,515,350,603 tokens
127,454,888,625 tokens
105,496,988,364 tokens
73,520,348,728 tokens
69,043,568,823 tokens
65,930,998,593 tokens
55,823,852,217 tokens
60,632,405,995 tokens
62,292,417,274 tokens
57,995,357,263 tokens
64,428,228,702 tokens
62,431,028,651 tokens
61,199,591,293 tokens
55,670,929,995 tokens
54,418,413,706 tokens
69,067,793,494 tokens
83,766,919,901 tokens
82,981,602,056 tokens
73,630,465,733 tokens
56,596,434,969 tokens
43,822,821,830 tokens
46,227,292,144 tokens
50,383,448,184 tokens
Daily observations · 43 days
| Date (UTC) | Reported tokens | Reported routes |
|---|---|---|
| Sep 7, 2026 | 50,383,448,184 | 1 |
| Sep 6, 2026 | 46,227,292,144 | 1 |
| Sep 5, 2026 | 43,822,821,830 | 1 |
| Sep 4, 2026 | 56,596,434,969 | 1 |
| Sep 3, 2026 | 73,630,465,733 | 1 |
| Sep 2, 2026 | 82,981,602,056 | 1 |
| Sep 1, 2026 | 83,766,919,901 | 1 |
| Aug 31, 2026 | 69,067,793,494 | 1 |
| Aug 30, 2026 | 54,418,413,706 | 1 |
| Aug 29, 2026 | 55,670,929,995 | 1 |
| Aug 28, 2026 | 61,199,591,293 | 1 |
| Aug 27, 2026 | 62,431,028,651 | 1 |
| Aug 26, 2026 | 64,428,228,702 | 1 |
| Aug 25, 2026 | 57,995,357,263 | 1 |
| Aug 24, 2026 | 62,292,417,274 | 1 |
| Aug 23, 2026 | 60,632,405,995 | 1 |
| Aug 22, 2026 | 55,823,852,217 | 1 |
| Aug 21, 2026 | 65,930,998,593 | 1 |
| Aug 20, 2026 | 69,043,568,823 | 1 |
| Aug 19, 2026 | 73,520,348,728 | 1 |
| Aug 18, 2026 | 105,496,988,364 | 1 |
| Aug 17, 2026 | 127,454,888,625 | 1 |
| Aug 16, 2026 | 113,515,350,603 | 1 |
| Aug 15, 2026 | 111,007,985,109 | 1 |
| Aug 14, 2026 | 125,957,921,257 | 1 |
| Aug 13, 2026 | 142,809,650,019 | 1 |
| Aug 12, 2026 | 130,992,558,780 | 1 |
| Aug 11, 2026 | 170,094,087,852 | 1 |
| Aug 10, 2026 | 171,499,623,423 | 1 |
| Aug 9, 2026 | 158,346,687,762 | 1 |
| Aug 8, 2026 | 164,102,567,452 | 1 |
| Aug 7, 2026 | 177,898,765,824 | 1 |
| Aug 6, 2026 | 183,760,329,848 | 1 |
| Aug 5, 2026 | 198,991,386,479 | 1 |
| Aug 4, 2026 | 188,156,210,411 | 1 |
| Aug 3, 2026 | 195,693,794,217 | 1 |
| Aug 2, 2026 | 174,724,791,518 | 1 |
| Aug 1, 2026 | 177,929,114,876 | 1 |
| Jul 31, 2026 | 202,112,542,879 | 1 |
| Jul 30, 2026 | 282,473,777,078 | 1 |
| Jul 29, 2026 | 324,874,967,863 | 1 |
| Jul 28, 2026 | 266,482,278,634 | 1 |
| Jul 27, 2026 | 227,222,837,235 | 1 |
Source: OpenRouter (openrouter.ai/rankings), as of Oct 24, 2026. CC BY 4.0. Totals include only individually published routes. Tokenizers differ by provider; the aggregated “other” category is never assigned to a model.
Technical details & API
Explore specifications, supported parameters, reasoning controls, and API identifiers for each OpenRouter route.
stepfun/step-3.7-flashObserved - Context
- 262,144
- Maximum output
- 230,400
- Tokenizer
- Other
- Input
- text, image, video
- Output
- text
- Moderation
- No
- Added to OpenRouter
- 2026-05-28
https://openrouter.ai/api/v1stepfun/step-3.7-flashstepfun/step-3.7-flash-20260528stepfun-ai/Step-3.7-FlashReasoning
- Required
- Yes
- Enabled by default
- Not reported
- Default effort
- medium
- Supported efforts
- high, medium, low
Supported parameters
frequency_penaltyinclude_reasoninglogprobsmax_tokenspresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
Frequently asked questions
Answers based on the model metadata and serving offers currently in the catalog.
How much does Step 3.7 Flash cost?
Listed rates start at $0.20 per million input tokens and $1.15 per million output tokens. Prices vary by provider and configuration; see Pricing for details.
Which providers offer Step 3.7 Flash?
There are 24 cataloged offers across 21 providers. Compare model IDs, prices and limits in Providers.
What is the context window?
The cataloged model context is 256,000 tokens. Each serving endpoint may apply a different limit.
Does it support tools and structured output?
Tool calling: Not reported. Structured output: Not reported. Support can vary by endpoint.
Resources
Catalog history · 3 recorded revisions
Metadata observations since this model was first cataloged. These are not model release versions.
- ·
5e53c0a86a37 - ·
8c9a397096a4 - ·
f976fca56afd