Qwen3 4B Thinking 2507

Qwen·Qwen/Qwen3-4B-Thinking-2507

Qwen3 4B Thinking 2507 is a language model published by Qwen. The latest OpenRouter endpoint refresh was unavailable, so no current serving claim is made. It accepts text input and returns text. The published context limit is 262K tokens. The catalogue declares support for tool use.

ModelCap Index · Current board

—

Not scored · 0% support · score interval not available

Market Gravity

11.8

Secondary market signal

OpenRouter endpoint status could not be refreshed. This model is not treated as currently served until a successful observation arrives.

Input

Free

per 1M tokens

Output

Free

per 1M tokens

Context

262K

tokens

Providers

endpoints

Gateway spend

—

latest · Vercel share

Providers

Uptime measured over the last 30 minutes

Current provider status is unavailable.

Hugging Face Inference Providers

Exact repository identity

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

Provider	Input	Output	Context	TTFT	Throughput	Features
nscale live	$0.01	$0.03	262K	593 ms	128.4 tok/s	tools · structured
featherless-ai live	—	—	—	—	—	—

Reported evaluation leads (3)

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 / task	Metric	Value	Revision	Provenance
Idavidrein/gpqa diamond	—	65.8	Unpinned `not published`	Reported Quarantined
MathArena/aime_2026 MathArena/aime_2026	—	82.5	Unpinned `not published`	Reported Quarantined
MathArena/hmmt_feb_2026 MathArena/hmmt_feb_2026	—	53.03	Unpinned `not published`	Reported Quarantined

ModelCap Index · Current board

Methodology

—

Not scored · excessive distance

Public rank basis: Not scored (abstained)
Index support: 0.0% · not applicable
Index interval: Not available
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/Qwen3-4B-Thinking-2507 · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata Qwen/Qwen3-4B-Thinking-2507@768f209d9ea81521153ed38c47d515654e938aea (not the evaluation revision)
Direct benchmark observations: 0
Architecture fallback: abstained · 0% coverage · 0 anchors

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: —
Evidence status: No admitted measured evidence
General preference: —
Coding: —
Agents & tools: —
Reasoning: —
Evidence breadth: 0%
Evidence coverage: 0%

Market and catalogue signals · secondary, never the language Index rank

Usage (OpenRouter popularity): —
Liquidity (providers × uptime): —
Open reach (HF downloads): 70.6
Surface (context / tools / modalities): 50.3
Freshness: 24.7

· HF 30d downloads 346275
· Catalogue: context 262144, tools
· First seen 2025-08-05

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

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.

Experimental capability estimate

Estimate policy

42.7

26.4–59.1 uncertainty interval

Excluded from rank

Confidence: 71%
Training support: 156 models · 94 lineages
Out-of-domain check: in domain
Feature coverage: 86%
Lineage-held-out error: 9.1 MAE

Experimental estimate. This layer predicts from admitted evidence and safe metadata only. It never enters ModelCap Score or rank.

Deployment readiness

Metadata index

62.7

partial · 70% metadata coverage

Not capability

Artifact reproducibility: 85.0
Access & legal clarity: 100.0
Deployability: 57.3
Evaluation provenance: 0.0

Missing: artifactReproducibility.base-lineage, deployability.measured-provider-uptime, evaluationProvenance.admitted-evaluation-observations, evaluationProvenance.capability-family-breadth, evaluationProvenance.evaluation-source-diversity, evaluationProvenance.identity-and-source-confidence, 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: 11.8
OpenRouter weekly popularity: Not listed

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

Awaiting evaluation

This model is visible because it is available in the live catalogue, but it has no trustworthy community result yet. ModelCap does not infer a quality score from price, features, or market activity.

Market Gravity breakdown

Usage55%: 0.0
Liquidity25%: 0.0
Open reach15%: 70.6
Freshness5%: 24.7

Specification

Context window: 262K tokens
Max output: —
Inputs: Text
Outputs: Text
Tokenizer: —
Cached input tokens: Not offered
First seen on OpenRouter: 5 Aug 2025

Capabilities

Not supported: Reasoning
Supported: Tool use
Not supported: Structured output
Not 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: 768f209d9ea8
Parameters: 4.0B
Architecture: Qwen3ForCausalLM
Model type: qwen3
Hugging Face downloads (30d): 346K
Downloads (all time): 6.8M
Likes: 608
Repository: Qwen/Qwen3-4B-Thinking-2507
Repository updated: 6 Aug 2025, 11:08 UTC