Skip to content
ModelCap

Qwen3.5-27B

All-versions archive #78Arena #129Open weights

Qwen·qwen/qwen3.5-27b

Qwen3.5-27B is a language model published by Qwen. 6 OpenRouter endpoints are published in the latest snapshot. It accepts text, image, video input and returns text. The published context limit is 262K tokens. The catalogue declares support for reasoning, tool use, structured outputs.

ModelCap Index · All-versions archive
64.7#78
Measured · 84% support · score interval 64.7–71.3
Market Gravity
46.8
Secondary market signal

Superseded by Qwen3.6 27B. It no longer holds a current-board position; its all-versions archive rank is #78.

Input
$0.195
per 1M tokens
Output
$1.56
per 1M tokens
Context
262K
tokens
Providers
6
endpoints
Gateway spend
—
latest · Vercel share

Providers

Uptime measured over the last 30 minutes
ProviderInputOutputContextQuantUptime
Alibabafp8Cheapest$0.195$1.56262Kfp899.93%
SiliconFlowfp8$0.25$2.00262Kfp897.02%
AtlasCloudfp8$0.27$2.16262Kfp899.69%
Novitabf16$0.30$2.40262Kbf16100.0%
Phala$0.30$2.40262K—99.42%
DeepInfrafp8$0.26$2.60262Kfp8100.0%

Hugging Face Inference Providers

Exact repository identity

Provider economics and telemetry published by the Hugging Face Router. These are operational signals, not benchmark results.

ProviderInputOutputContextTTFTThroughputFeatures
novita
live
$0.30$2.40262K995 ms72.4 tok/stools
deepinfra
live
$0.26$2.60262K613 ms52.4 tok/stools · structured
featherless-ai
live
——————

Reported evaluation leads (11)

Excluded from score

Structured rows reported in this exact model repository at the pinned revision below. They are useful research leads, but remain quarantined until dataset revision, harness, configuration and identity pass ModelCap’s independent admission review.

Dataset / taskMetricValueRevisionProvenance
cais/hle
hle
—24.3
Unpinned
not published
Reported
Quarantined
cais/hle
hle
—48.5
Unpinned
not published
Reported
Quarantined
harborframework/terminal-bench-2.0
terminalbench_2
—41.6
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—81.8182
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—85.5
Unpinned
not published
Reported
Quarantined
likaixin/ScreenSpot-Pro
overall
—70.3
Unpinned
not published
Reported
Quarantined
MathArena/aime_2026
MathArena/aime_2026
—90.83
Unpinned
not published
Reported
Quarantined
MathArena/hmmt_feb_2026
MathArena/hmmt_feb_2026
—81.06
Unpinned
not published
Reported
Quarantined
MMMU/MMMU_Pro
mmmu_pro_vision
—75
Unpinned
not published
Reported
Quarantined
SWE-bench/SWE-bench_Verified
swe_bench_%_resolved
—72.4
Unpinned
not published
Reported
Quarantined
TIGER-Lab/MMLU-Pro
mmlu_pro
—86.1
Unpinned
not published
Reported
Quarantined

Other versions (1)

ModelCap Index · All-versions archive

Methodology
64.7
Measured · 2 public benchmark observations across 2 boards
All-versions archive
#78
Public rank basis
Measured (measured)
Index support
83.5% · strong
Index interval
64.7–71.3
Rank posterior
Not available
Top 5 / Top 10 probability
Not available / Not available
Posterior as of
Snapshot timestamp unavailable
Identity binding
Exact catalogue product · qwen/qwen/qwen3.5-27b · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata Qwen/Qwen3.5-27B@fc05daec18b0a78c049392ed2e771dde82bdf654 (not the evaluation revision)
Direct benchmark observations
2
Breadth-adjusted BES point estimate
68.0
Bayesian Evidence Score · measured evidence input only

BES normalizes admitted public benchmark evidence for measured Index rows. Its legacy score and observables prior do not define the public language rank.

Observed capability
68.0
Evidence status
confirmed · 83% mass
General preference
68.0
Coding
67.8
Agents & tools
—
Reasoning
—
Evidence breadth
92%
Evidence coverage
84%
Market and catalogue signals · secondary, never the language Index rank
Usage (OpenRouter popularity)
14.3
Liquidity (providers × uptime)
89.4
Open reach (HF downloads)
92.9
Surface (context / tools / modalities)
90.3
Freshness
53.8
  • · OpenRouter popularity #132
  • · 6 live providers
  • · HF 30d downloads 2563438
  • · Catalogue: context 262144, reasoning, tools, modalities Text/Image/Video
  • · First seen 2026-02-25

These adoption and deployment observations remain context only. They do not change this model's ModelCap Index score or rank.

BES evidence selected · measured input only

The bounded 0–100 capability score blends each source’s competitive placement with its published achievement, then combines capability families. Evidence breadth adds a modest uncertainty adjustment; price and popularity are not part of benchmark-led rank.

Deployment readiness

Metadata index
85.4
strong · 94% metadata coverage
Not capability
Artifact reproducibility
85.0
Access & legal clarity
100.0
Deployability
97.7
Evaluation provenance
53.0

Missing: artifactReproducibility.base-lineage, evaluationProvenance.evaluation-dataset-revisions.

A metadata completeness and deployability index—not a safety certification, quality grade, or production-readiness claim.

Market activity

Venue coverage
Market Gravity
46.8
OpenRouter weekly popularity
#132 of 298

Market Gravity uses OpenRouter’s full-catalogue popularity order. Sparse Vercel top-ten observations appear only when published and do not affect the score.

Arena preference

Status
Ranked
Source
Arena
Official rank
#129 of 385
Rating
1407.9
95% confidence interval
1404–1412
Votes
27K
Source snapshot
6 Aug 2026, 13:00 UTC
Category
overall
Identity match confidence
100%

Market Gravity breakdown

Usage55%
14.3

OpenRouter weekly popularity across the full model catalogue.

Liquidity25%
89.4

Independent providers versus the model's open or closed cohort, weighted by uptime.

Open reach15%
92.9

Hugging Face 30-day downloads for open models; neutral for closed models.

Freshness5%
53.8

Time since first public availability, on a six-month half-life.

Specification

Context window
262K tokens
Max output
66K tokens
Inputs
Text, Image, Video
Outputs
Text
Tokenizer
Qwen3
Cached input tokens
Not offered
First seen on OpenRouter
25 Feb 2026

Capabilities

  • Supported: Reasoning
  • Supported: Tool use
  • Supported: Structured output
  • Supported: Response format
  • Not supported: Moderated

Weights & access

Open weights. Downloadable weights are published under an OSI or free-culture license in the current repository metadata. This describes weight availability, not whether the full training stack qualifies as Open Source AI.

Access
Open weights
Licence
apache-2.0
Pinned revision
fc05daec18b0
Parameters
27.8B
Architecture
Qwen3_5ForConditionalGeneration
Model type
qwen3_5
Hugging Face downloads (30d)
2.6M
Downloads (all time)
15.9M
Likes
1K
Repository
Qwen/Qwen3.5-27B
Repository updated
24 Apr 2026, 02:54 UTC