Skip to content
AI Atlas
ModelActiveOpen weightsIdentity probable

Step 3.7 Flash

StepFunfamily · Step3.7huggingface.co/stepfun-ai/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...

Open in Graph
data quality72

Updated 4 h ago · first seen 11 Sept 2026

model_01M294WVXWCJK3P3D3PV5X7Q3G

Overview

Identity

Canonical model
Yesidentity confidence: mediumOne row per real model release. Artifacts (checkpoints, quantisations, conversions) and folded evaluation variants point here.
Official checkpoints
official_checkpoints = hf_repo identifiers carried by the model itself; artifacts are separate entities pointing here through canonical_id.
Artifacts
3 quantizations0 official · 3 third-partySeparate entities (checkpoint · quantization · conversion · packaging) pointing to this model through canonical_id.
Provider deployments
1
API aliases
step-3-7-flashstepfun/step-3.7-flashIdentifiers under which providers and evaluators refer to this model.
Folded evaluation variants
0Effort / thinking variants (…-high, …-non-reasoning) are result configurations of this model, not separate models. Their old URLs redirect here.

Openness

Open weights— weights downloadable under Apache-2.0; commercial use allowed; redistribution allowed; derivatives allowed; 4 dimensions unknown.

Weights downloadable under a permissive or Creative Commons licence allowing commercial use; code or data may be missing.

  • Weights

    Yes

  • Inference code

    —

  • Training code

    —

  • Training data

    —

  • Dataset

    —

  • Commercial use

    Yes

  • Redistribution

    Yes

  • Derivatives

    Yes

Licence: Apache License 2.0 (permissive · SPDX Apache-2.0)

dimensions marked null are unknown, not false

Key facts

Release date

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Openrouter id

Source:OpenRouter public model & pricing listingT2observed 5 d agomedium

Architecture

Architecture

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Model type

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Parameters

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Weights dtype

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

File size

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Library name

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Pipeline tag

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Hugging Face repo

Source:OpenRouter public model & pricing listingT2observed 5 d agomedium

Capabilities

Modalities

Modalities
imagetextvideo
Input
imagetextvideo
Output
text

Capabilities

  • Tool calling

    Yes

    OpenRouter public model & pricing listing · T2

  • Structured output

    Yes

    OpenRouter public model & pricing listing · T2

  • Reasoning

    Yes

    OpenRouter public model & pricing listing · T2

  • Vision

    Yes

    OpenRouter public model & pricing listing · T2

  • Audio

    Unavailable

  • Fine-tuning available

    Unavailable

Context window

Source:OpenRouter public model & pricing listingT2observed 5 d agomedium

Max output

Source:OpenRouter public model & pricing listingT2observed 5 d agomedium

Languages

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 4 d agomedium

Benchmarks18

Compare with another model →

Comparable same task and conditions · Partially comparable same task, conditions differ (effort, temperature, judge) · Not comparable different variant or metric

Benchmark results grouped by comparability group
Benchmark · groupBest scoreTrustConfigurationResultsvs leaderEvaluatedSource
Terminal-Benchagentic · accuracy · variant=hard · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisvarianthardreasoningonconditions differ across rows → partially comparable2−30.3 ptvs gpt-5.6-solobs. 16 Sept 2026artificialanalysis.aiT2
Terminal-Benchagentic · accuracy · variant=v2.1 · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisvariantv2.1reasoningonconditions differ across rows → partially comparable2non-primary groupobs. 15 Sept 2026artificialanalysis.aiT2
τ²-benchagentic · pass^1 · variant=Telecom · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantTelecomreasoningonconditions differ across rows → partially comparable1non-primary groupobs. 16 Sept 2026artificialanalysis.aiT2
τ²-benchagentic · pass^1 · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysis1−0.58 ptvs Z.ai GLM 5.2obs. 11 Sept 2026artificialanalysis.aiT2
SciCodecoding · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoningonconditions differ across rows → partially comparable2−19.2 ptvs Claude Fable 5.1obs. 16 Sept 2026artificialanalysis.aiT2
Artificial Analysis Intelligence Indexcomposite · indexIndependentreasoningonversion4.3conditions differ across rows → partially comparable2−33.9vs Claude Fable 5.1obs. 16 Sept 2026artificialanalysis.aiT2
IFBenchinstruction-following · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoningonconditions differ across rows → partially comparable2−16.0 ptvs Grok 4.3obs. 16 Sept 2026artificialanalysis.aiT2
Humanity's Last Examknowledge · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoningonconditions differ across rows → partially comparable2−37.7 ptvs Claude Fable 5.1obs. 16 Sept 2026artificialanalysis.aiT2
MMMU-Promultimodal · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoningonconditions differ across rows → partially comparable2−11.6 ptvs gpt-6-astraobs. 16 Sept 2026artificialanalysis.aiT2
GPQA Diamondreasoning · accuracy · variant=Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantDiamondreasoningonconditions differ across rows → partially comparable1non-primary groupobs. 16 Sept 2026artificialanalysis.aiT2
GPQA Diamondreasoning · accuracy · variant=GPQA Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantGPQA Diamond1−15.4 ptvs gpt-6-astraobs. 11 Sept 2026artificialanalysis.aiT2

Current rows only, grouped by benchmark → canonical metric → comparability group (task configuration). Effort variants folded into this model appear as rows of the same group. 18 current rows in total. “vs leader” compares with the current leader of the benchmark's primary group only; other groups are not directly comparable. Comparability rules →

Providers & Pricing1

All offers in the price terminal →
Provider deployments, cheapest output first
ProviderContextInput / 1MCached inOutput / 1MStatusObservedSource
OpenRoutercheapest outputstepfun/step-3.7-flash256Kout 230.4K$0.04active2 h agosince 11 Sept 2026openrouter.aiT2

USD per 1M tokens as published by each provider; native units (per-request fees, flex/priority tiers) are kept verbatim. Rows are append-only — every price change is kept in the history below. Cost of a workload →

Price history

Step lines per provider; amber markers are recorded changes. Click a marker or a row for the evidence behind that price.

Output price · USD / 1M tokens 1 provider

Output price history of Step 3.7 Flash$0$0.50$1$1.5Sept 26Sept 26Sept 26Sept 26Sept 26OpenRouter: first observed → $1.15 · 11 Sept 2026
  • OpenRouter
  • OpenRouterfirst observed $1.1511 Sept 2026

Input price · USD / 1M tokens 1 provider

Input price history of Step 3.7 Flash$0$0.05$0.10$0.15$0.20$0.25Sept 26Sept 26Sept 26Sept 26Sept 26OpenRouter: first observed → $0.20 · 11 Sept 2026
  • OpenRouter
  • OpenRouterfirst observed $0.2011 Sept 2026

Hardware fit37

Estimated

15 of 37 device × quantization combinations fit.

Run locally: your machine →
Estimated hardware fit
HardwareQuantizationDevice memoryEst. memoryFits
Apple M3 Ultra4bit—116.3 GB est.Yes
Mac Studio (Apple M5 Ultra)4bit—116.3 GB est.Yes
Apple M2 Ultra4bit—116.3 GB est.Yes
Apple M1 Ultra4bit—116.3 GB est.Yes
Apple M3 Max4bit—116.3 GB est.Yes
Apple M4 Max4bit—116.3 GB est.Yes
Mac Studio (Apple M5 Max)4bit—116.3 GB est.Yes
MacBook Pro (Apple M5 Max)4bit—116.3 GB est.Yes
NVIDIA DGX Spark4bit128 GB116.3 GB est.Yes
NVIDIA H2004bit141 GB116.3 GB est.Yes
NVIDIA H200 NVL4bit141 GB116.3 GB est.Yes
NVIDIA B2004bit180 GB116.3 GB est.Yes
AMD Instinct MI300X4bit192 GB116.3 GB est.Yes
AMD Instinct MI325X4bit256 GB116.3 GB est.Yes
NVIDIA DGX B2004bit1,440 GB116.3 GB est.Yes
Apple M2 Max4bit—116.3 GB est.No
Apple M1 Max4bit—116.3 GB est.No
Apple M4 Pro4bit—116.3 GB est.No
Mac mini (Apple M5 Pro)4bit—116.3 GB est.No
MacBook Pro (Apple M5 Pro)4bit—116.3 GB est.No
Apple M3 Pro4bit—116.3 GB est.No
Apple M1 Pro4bit—116.3 GB est.No
Apple M2 Pro4bit—116.3 GB est.No
Apple M44bit—116.3 GB est.No
iMac (Apple M4)4bit—116.3 GB est.No
Mac mini (Apple M6)4bit—116.3 GB est.No
MacBook Air (Apple M5)4bit—116.3 GB est.No
MacBook Pro (Apple M5)4bit—116.3 GB est.No
Apple M24bit—116.3 GB est.No
Apple M34bit—116.3 GB est.No
Apple M14bit—116.3 GB est.No
NVIDIA GeForce RTX 30904bit24 GB116.3 GB est.No
NVIDIA GeForce RTX 40904bit24 GB116.3 GB est.No
NVIDIA GeForce RTX 50904bit32 GB116.3 GB est.No
NVIDIA A100 80GB4bit80 GB116.3 GB est.No
NVIDIA H100 SXM4bit80 GB116.3 GB est.No
NVIDIA H100 NVL4bit94 GB116.3 GB est.No
Assumptions (7)
  • Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded).
  • bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
  • KV cache: 2 × layers × kv_heads × head_dim × 2 bytes × context × batch when the architecture is known; otherwise 0.5 GB per 8 192 tokens (× batch), independent of architecture (GQA/MLA models need less).
  • A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
  • Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
  • Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).
  • Multi-GPU: device memories are summed; interconnect bandwidth, tensor-parallel replication and pipeline bubbles are not modelled.

Lineage

Open in Graph →
Explicit derived_from / fine_tuned_from / distilled_from relations stated by sources; artifacts collapsed by kind.
DESCENDANTS 0 · ARTIFACTS3 quantizationsartifacts · collapsed3 quantizations — artifacts · collapsedStep 3.7 Flash201.4B params · this modelStep 3.7 Flash — 201.4B params · this model

    Versions & Artifacts3

    Version history

    Context windowfirst observation only

    11 Sept 2026current

    Licensefirst observation only

    12 Sept 2026current

    Max outputfirst observation only

    11 Sept 2026current

    Opennessfirst observation only

    11 Sept 2026current

    Parametersfirst observation only

    12 Sept 2026current

    Each hop is a claim: click a value for its source, tier and observation time. Nothing is overwritten — a new observation closes the previous claim.

    Artifacts 3

    quantizations 3

    Timeline14

    Full timeline →

    Change history

    Temporal, append-only claims: a new observation closes the previous claim instead of overwriting it. Rewind the record with the as-of picker.
    1 claims · 1 propertiesShow all properties

    Library namelibrary_name1

    Claim history for Library name
    ValueValid from → toStatusSourceConfidenceExtractor
    transformers → currentcurrentHugging Face Hub (public pages, model cards, papers)T2mediumdeterministic

    Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →

    Provenance

    Attributed facts

    43

    Source tiers

    T243

    Freshest observation

    4 h ago

    Conflicts

    None

    Source documents 5

    Source documents
    SourceDocumentTypeTierLast observedSnapshots
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/stepfun-ai/Step-3.7-Flash-GGUF model_pageT2· Quality secondary9 min ago25
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/stepfun-ai/Step-3.7-Flash model_pageT2· Quality secondary9 min ago25
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/models?author=stepfun-ai&p=0&sort=downloads listingT2· Quality secondary46 min ago26
    OpenRouter public model & pricing listingopenrouter.ai/api/v1/models catalogueT2· Quality secondary2 h ago75
    Artificial Analysisartificialanalysis.ai/leaderboards/models leaderboardT2· Quality secondary6 h ago17

    Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.

    Data quality (72/100) measures how well AI Atlas knows this entity — completeness, primary-source ratio, freshness, conflicts — never how good the model is. Methodology →