Skip to content
AI Atlas
ModelActiveOpen weightsIdentity probable

Qwen3.8 27B

Qwenfamily · Qwen3.8huggingface.co/Qwen/Qwen3.8-27B

Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...

Open in Graph
data quality73

Updated 6 h ago · first seen 11 Sept 2026

model_01M294WVR4JWTTVTSEYX2E0VCE

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
9 quantizations0 official · 8 third-partySeparate entities (checkpoint · quantization · conversion · packaging) pointing to this model through canonical_id.
Provider deployments
2
API aliases
qwen/qwen3.8-27bqwen3-8-27bqwen3-8-27b-lowqwen3-8-27b-mediumqwen3-8-27b-non-reasoningIdentifiers under which providers and evaluators refer to this model.
Folded evaluation variants
3Effort / 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 · stated as “apache-2.0”)

dimensions marked null are unknown, not false

Key facts

Release date

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 11 h agomedium

Model card
huggingface.co/Qwen/Qwen3.8-27B

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 16 h agomedium

Openrouter id

Source:OpenRouter public model & pricing listingT2observed 16 h agomedium

Architecture

Architecture

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 15 h agomedium

Model type

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 15 h agomedium

Parameters

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 16 h agomedium

Tokenizer

Source:OpenRouter public model & pricing listingT2observed 16 h agomedium

Weights dtype

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 15 h agomedium

File size

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 15 h agomedium

Library name

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 15 h agomedium

Pipeline tag

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 16 h agomedium

Hugging Face repo

Source:OpenRouter public model & pricing listingT2observed 16 h 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 13 h agomedium

Max output

Source:OpenRouter public model & pricing listingT2observed 16 h agomedium

Tokenizer

Source:OpenRouter public model & pricing listingT2observed 16 h agomedium

Benchmarks62

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=v4.0 · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisvariantv4.0reasoning_effortxhighconditions differ across rows → partially comparable6non-primary groupobs. 12 Sept 2026artificialanalysis.aiT2
Terminal-Benchagentic · accuracy · variant=v2.1 · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisvariantv2.1reasoning_effortxhighconditions differ across rows → partially comparable8non-primary groupobs. 12 Sept 2026artificialanalysis.aiT2
LiveBench Agentic Codingcoding · average score · variant=Agentic CodingOfficial boardvariantAgentic Codingreasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−15.9 ptvs DeepSeek-V4.1-Flash25 Jun 2026livebench.aiT2
LiveBench Codingcoding · average score · variant=CodingOfficial boardvariantCodingreasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−10.7 ptvs Claude Fable 5.125 Jun 2026livebench.aiT2
SciCodecoding · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoning_effortxhighconditions differ across rows → partially comparable8−16.4 ptvs Claude Fable 5.1obs. 12 Sept 2026artificialanalysis.aiT2
Artificial Analysis Intelligence Indexcomposite · indexIndependentreasoning_effortxhighversion4.3conditions differ across rows → partially comparable8−19.5vs Claude Fable 5.1obs. 12 Sept 2026artificialanalysis.aiT2
LiveBenchgeneral · global_averageOfficial boardreasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−8.14 ptvs Claude Fable 5.125 Jun 2026livebench.aiT2
LiveBench Data Analysisgeneral · average score · variant=Data AnalysisOfficial boardvariantData Analysisreasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−6.39 ptvs gpt-6-astra25 Jun 2026livebench.aiT2
LiveBench Languagegeneral · average score · variant=LanguageOfficial boardvariantLanguagereasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−16.3 ptvs Claude Fable 525 Jun 2026livebench.aiT2
LiveBench Instruction Followinginstruction-following · average score · variant=IFOfficial boardvariantIFreasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−8.75 ptvs Gemini 3.8 Flash25 Jun 2026livebench.aiT2
Humanity's Last Examknowledge · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoning_effortxhighconditions differ across rows → partially comparable8−25.2 ptvs Claude Fable 5.1obs. 12 Sept 2026artificialanalysis.aiT2
LiveBench Mathematicsmath · average score · variant=MathematicsOfficial boardvariantMathematicsreasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−10.8 ptvs Claude Fable 5.125 Jun 2026livebench.aiT2
MMMU-Promultimodal · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoning_effortxhighconditions differ across rows → partially comparable8−10.6 ptvs gpt-6-astraobs. 12 Sept 2026artificialanalysis.aiT2
GPQA Diamondreasoning · accuracy · variant=Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantDiamondreasoning_effortxhighconditions differ across rows → partially comparable4non-primary groupobs. 12 Sept 2026artificialanalysis.aiT2
GPQA Diamondreasoning · accuracy · variant=GPQA Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantGPQA Diamond4−5.75 ptvs gpt-6-astraobs. 11 Sept 2026artificialanalysis.aiT2
LiveBench Reasoningreasoning · average score · variant=ReasoningOfficial boardvariantReasoningreasoningoffrelease2026-06-25conditions differ across rows → partially comparable1−12.6 ptvs gpt-6-astra25 Jun 2026livebench.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. 62 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 & Pricing2

All offers in the price terminal →
Provider deployments, cheapest output first
ProviderContextInput / 1MCached inOutput / 1MNative unitsStatusObservedSource
OpenRoutercheapest outputqwen/qwen3.8-27b262.1Kout 131.1K$0.15—active2 h agosince 12 Sept 2026openrouter.aiT2
GroqCloudqwen/qwen3.8-27b131Kout 16.4K—output_tokens_per_second=450active3 h agosince 11 Sept 2026console.groq.comT1

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 2 providers

Output price history of Qwen3.8 27B$0$1$2$3$4$5Sept 26Sept 26Sept 26Sept 26Sept 26OpenRouter: first observed → $3 · 11 Sept 2026GroqCloud: first observed → $4 · 11 Sept 2026OpenRouter: $3 → $2.55 · 12 Sept 2026
  • OpenRouter
  • GroqCloud
  • OpenRouter$3 → $2.5512 Sept 2026
  • GroqCloudfirst observed $411 Sept 2026
  • OpenRouterfirst observed $311 Sept 2026

Input price · USD / 1M tokens 2 providers

Input price history of Qwen3.8 27B$0$0.20$0.40$0.60$0.80$1Sept 26Sept 26Sept 26Sept 26Sept 26OpenRouter: first observed → $0.42 · 11 Sept 2026GroqCloud: first observed → $0.80 · 11 Sept 2026OpenRouter: $0.42 → $0.214 · 12 Sept 2026
  • OpenRouter
  • GroqCloud
  • OpenRouter$0.42 → $0.21412 Sept 2026
  • GroqCloudfirst observed $0.8011 Sept 2026
  • OpenRouterfirst observed $0.4211 Sept 2026

Hardware fit37

Estimated

36 of 37 device × quantization combinations fit.

Run locally: your machine →
Estimated hardware fit
HardwareQuantizationDevice memoryEst. memoryFits
Apple M3 Ultra4bit—16.5 GB est.Yes
Mac Studio (Apple M5 Ultra)4bit—16.5 GB est.Yes
Apple M2 Ultra4bit—16.5 GB est.Yes
Apple M1 Ultra4bit—16.5 GB est.Yes
Apple M3 Max4bit—16.5 GB est.Yes
Apple M4 Max4bit—16.5 GB est.Yes
Mac Studio (Apple M5 Max)4bit—16.5 GB est.Yes
MacBook Pro (Apple M5 Max)4bit—16.5 GB est.Yes
Apple M2 Max4bit—16.5 GB est.Yes
Apple M1 Max4bit—16.5 GB est.Yes
Apple M4 Pro4bit—16.5 GB est.Yes
Mac mini (Apple M5 Pro)4bit—16.5 GB est.Yes
MacBook Pro (Apple M5 Pro)4bit—16.5 GB est.Yes
Apple M3 Pro4bit—16.5 GB est.Yes
Apple M1 Pro4bit—16.5 GB est.Yes
Apple M2 Pro4bit—16.5 GB est.Yes
Apple M44bit—16.5 GB est.Yes
iMac (Apple M4)4bit—16.5 GB est.Yes
Mac mini (Apple M6)4bit—16.5 GB est.Yes
MacBook Air (Apple M5)4bit—16.5 GB est.Yes
MacBook Pro (Apple M5)4bit—16.5 GB est.Yes
Apple M24bit—16.5 GB est.Yes
Apple M34bit—16.5 GB est.Yes
NVIDIA GeForce RTX 30904bit24 GB16.5 GB est.Yes
NVIDIA GeForce RTX 40904bit24 GB16.5 GB est.Yes
NVIDIA GeForce RTX 50904bit32 GB16.5 GB est.Yes
NVIDIA A100 80GB4bit80 GB16.5 GB est.Yes
NVIDIA H100 SXM4bit80 GB16.5 GB est.Yes
NVIDIA H100 NVL4bit94 GB16.5 GB est.Yes
NVIDIA DGX Spark4bit128 GB16.5 GB est.Yes
NVIDIA H2004bit141 GB16.5 GB est.Yes
NVIDIA H200 NVL4bit141 GB16.5 GB est.Yes
NVIDIA B2004bit180 GB16.5 GB est.Yes
AMD Instinct MI300X4bit192 GB16.5 GB est.Yes
AMD Instinct MI325X4bit256 GB16.5 GB est.Yes
NVIDIA DGX B2004bit1,440 GB16.5 GB est.Yes
Apple M14bit—16.5 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 · ARTIFACTS9 quantizationsartifacts · collapsed9 quantizations — artifacts · collapsedQwen3.8 27B27.8B params · this modelQwen3.8 27B — 27.8B params · this model

    Versions & Artifacts9

    Version history

    Context window2 changes

    11 Sept 2026→11 Sept 2026→12 Sept 2026current

    Licensefirst observation only

    11 Sept 2026current

    Max outputfirst observation only

    11 Sept 2026current

    Opennessfirst observation only

    11 Sept 2026current

    Parametersfirst observation only

    11 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 9

    quantizations 9

    Timeline30

    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

    Derivatives allowedderivatives_allowed1

    Claim history for Derivatives allowed
    ValueValid from → toStatusSourceConfidenceExtractor
    Yes → currentcurrentAI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2highderived

    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

    45

    Source tiers

    T1T22 / 43

    Freshest observation

    6 h ago

    Conflicts

    None

    Source documents 18

    Source documents
    SourceDocumentTypeTierLast observedSnapshots
    Groq — docs & pricingconsole.groq.com/docs/models model_docsT1· Official3 h ago5
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/unsloth/Qwen3.8-27B-NVFP4 model_pageT2· Quality secondary56 min ago5
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/unsloth/Qwen3.8-27B-GGUF model_pageT2· Quality secondary56 min ago5
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/Qwen/Qwen3.8-27B model_pageT2· Quality secondary1 h ago5
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/Qwen/Qwen3.8-27B-FP8 model_pageT2· Quality secondary1 h ago6
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/perplexity-ai/pplx-qwen-3-8-27b-dflash2-20260819 model_pageT2· Quality secondary1 h ago6
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/perplexity-ai/pplx-computer-qwen-3-8-27b-dflash2-20260824 model_pageT2· Quality secondary1 h ago5
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/perplexity-ai/pplx-computer-qwen-3-8-27b-dflash2-gguf-20260826 model_pageT2· Quality secondary1 h ago5
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/mlx-community/Qwen3.8-27B-8bit model_pageT2· Quality secondary1 h ago6
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/mlx-community/Qwen3.8-27B-4bit model_pageT2· Quality secondary1 h ago5
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/amd/Qwen3.8-27B-Quark-AWQ-MXFP4 model_pageT2· Quality secondary2 h ago7
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/amd/Qwen3.8-27B-Quark-AWQ-INT4-W4A16 model_pageT2· Quality secondary2 h ago7
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/models?author=Qwen&p=0&sort=downloads listingT2· Quality secondary2 h ago8
    OpenRouter public model & pricing listingopenrouter.ai/api/v1/models catalogueT2· Quality secondary2 h ago11
    Artificial Analysisartificialanalysis.ai/leaderboards/models leaderboardT2· Quality secondary6 h ago2
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/Qwen/Qwen3.8-27B/raw/main/README.md model_cardT2· Quality secondary9 h ago1
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/bartowski/Qwen3.8-27B-GGUF/raw/main/README.md model_cardT2· Quality secondary9 h ago1
    LiveBenchlivebench.ai/table_2026_06_25.csv leaderboardT2· Quality secondary14 h ago1

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

    Data quality (73/100) measures how well AI Atlas knows this entity — completeness, primary-source ratio, freshness, conflicts — never how good the model is. Methodology →