AlibabaLanguage model
QwQ 32B
QwQ 32B has no overall index value yet because there are too few independent benchmark results.
PopularityRank 176 of 785
45out of 100same as last month
based on search, usage and attention
Summary
QwQ 32B is a reasoning specialist in the Qwen family. Unlike a conventional instruct model, it was trained with supervised examples and reinforcement learning to work through difficult tasks step by step.
The open model has roughly 32.5 billion parameters and a full context length of 131,072 tokens. Prompts longer than 8,192 tokens require YaRN to be enabled.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 101 published benchmark results
Page 1 of 9
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
IFBench: Overall (Show measurement, test conditions, and source)
- Source value
- 38.7755102040816
- Score
- 38.7755102040816
- Metric
- IFBench prompt-level loose accuracy (pass@1)
- Unit
- %
- Benchmark version
- AA test set: 294 questions
- Category
- general
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- superseded
- Retrieved at
- 2026-11-02
- Methodology
- Exact AA configuration: QwQ 32B (release qwq-32b, row qwq-32b, UUID ceb4d610-d0a4-48c1-bea0-80ed76f1e5ca, reasoning=true, effort=not specified, suffix=not specified). AA source value for average reasoning tokens before the answer: 2491. Its methodology averages 60 varied prompts and assumes 2,000 when no average is available. This is not a token cap. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Source model ID: artificial-analysis/ceb4d610-d0a4-48c1-bea0-80ed76f1e5ca/qwq-32b/qwq-32b. Source model name: QwQ 32B. Provider reported by source: Alibaba. Model snapshot: QwQ 32B. Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: reasoning enabled. Score display: 38.8%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-11-02-67d46be20a5b. Source payload hash: 67d46be20a5b412db6cfb60ad190441fcaa5513500b8fd47d033168d0e97b6b2. Parser version: artificial-analysis-rsc-models-v3. Imported at: 2026-11-02. Source cohort: artificial-analysis-ifbench:AA test set: 294 questions:standalone:reasoning:effort=unspecified:suffix=unspecified
- Comparison methodology
- Exact AA configuration: QwQ 32B (release qwq-32b, row qwq-32b, UUID ceb4d610-d0a4-48c1-bea0-80ed76f1e5ca, reasoning=true, effort=not specified, suffix=not specified). AA source value for average reasoning tokens before the answer: 2491. Its methodology averages 60 varied prompts and assumes 2,000 when no average is available. This is not a token cap. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: reasoning enabled
- Context
- The often cited 58 refers to verifiable out-of-domain constraint types, not the number of questions in AA's test run.. This result is retained as historical evidence and is not interchangeable with current results.
IFBench: Overall (Show measurement, test conditions, and source)
- Source value
- 38.7755102040816
- Score
- 38.7755102040816
- Metric
- IFBench prompt-level loose accuracy (pass@1)
- Unit
- %
- Benchmark version
- AA test set: 294 questions
- Category
- general
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-11-07
- Methodology
- Exact AA configuration: QwQ 32B (release qwq-32b, row qwq-32b, UUID ceb4d610-d0a4-48c1-bea0-80ed76f1e5ca, reasoning=true, effort=not specified, suffix=not specified). This source capture does not provide an average reasoning-token count. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Source model ID: artificial-analysis/ceb4d610-d0a4-48c1-bea0-80ed76f1e5ca/qwq-32b/qwq-32b. Source model name: QwQ 32B. Provider reported by source: Alibaba. Model snapshot: QwQ 32B. Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: reasoning enabled. Score display: 38.8%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-11-07-b96e3ffc8c5b. Source payload hash: b96e3ffc8c5b052aa56f96d05e89d02b8cc234a75085820fca0ca2a39e498274. Parser version: artificial-analysis-rsc-models-v4. Imported at: 2026-11-07. Source cohort: artificial-analysis-ifbench:AA test set: 294 questions:standalone:reasoning:effort=unspecified:suffix=unspecified
- Comparison methodology
- Exact AA configuration: QwQ 32B (release qwq-32b, row qwq-32b, UUID ceb4d610-d0a4-48c1-bea0-80ed76f1e5ca, reasoning=true, effort=not specified, suffix=not specified). This source capture does not provide an average reasoning-token count. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: reasoning enabled
- Context
- The often cited 58 refers to verifiable out-of-domain constraint types, not the number of questions in AA's test run.
LongBench v2 (Show measurement, test conditions, and source)
- Source value
- 48.9
- Score
- 48.9
- Metric
- Overall, w/ CoT
- Unit
- %
- Category
- Overall
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- LongBench v2 authors, official leaderboard
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact LongBench v2 row QwQ-32B; context column 128k. The source columns w/o CoT and w/ CoT are recorded separately and only where a numeric cell exists. The source-wide YaRN factor-4 note is retained; the leaderboard does not state a per-row 16K thinking budget for this non-Qwen3 row.
- Context
- LongBench v2 does not publish an immutable inference checkpoint/commit, task denominator, or confidence interval for this row.
LongBench v2 (Show measurement, test conditions, and source)
- Source value
- 58.9
- Score
- 58.9
- Metric
- Easy, w/ CoT
- Unit
- %
- Category
- Easy
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- LongBench v2 authors, official leaderboard
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact LongBench v2 row QwQ-32B; context column 128k. The source columns w/o CoT and w/ CoT are recorded separately and only where a numeric cell exists. The source-wide YaRN factor-4 note is retained; the leaderboard does not state a per-row 16K thinking budget for this non-Qwen3 row.
- Context
- LongBench v2 does not publish an immutable inference checkpoint/commit, task denominator, or confidence interval for this row.
LongBench v2 (Show measurement, test conditions, and source)
- Source value
- 42.8
- Score
- 42.8
- Metric
- Hard, w/ CoT
- Unit
- %
- Category
- Hard
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- LongBench v2 authors, official leaderboard
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact LongBench v2 row QwQ-32B; context column 128k. The source columns w/o CoT and w/ CoT are recorded separately and only where a numeric cell exists. The source-wide YaRN factor-4 note is retained; the leaderboard does not state a per-row 16K thinking budget for this non-Qwen3 row.
- Context
- LongBench v2 does not publish an immutable inference checkpoint/commit, task denominator, or confidence interval for this row.
LongBench v2 (Show measurement, test conditions, and source)
- Source value
- 54.4
- Score
- 54.4
- Metric
- Short, w/ CoT
- Unit
- %
- Category
- Short
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- LongBench v2 authors, official leaderboard
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact LongBench v2 row QwQ-32B; context column 128k. The source columns w/o CoT and w/ CoT are recorded separately and only where a numeric cell exists. The source-wide YaRN factor-4 note is retained; the leaderboard does not state a per-row 16K thinking budget for this non-Qwen3 row.
- Context
- LongBench v2 does not publish an immutable inference checkpoint/commit, task denominator, or confidence interval for this row.
LongBench v2 (Show measurement, test conditions, and source)
- Source value
- 44.7
- Score
- 44.7
- Metric
- Medium, w/ CoT
- Unit
- %
- Category
- Medium
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- LongBench v2 authors, official leaderboard
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact LongBench v2 row QwQ-32B; context column 128k. The source columns w/o CoT and w/ CoT are recorded separately and only where a numeric cell exists. The source-wide YaRN factor-4 note is retained; the leaderboard does not state a per-row 16K thinking budget for this non-Qwen3 row.
- Context
- LongBench v2 does not publish an immutable inference checkpoint/commit, task denominator, or confidence interval for this row.
LongBench v2 (Show measurement, test conditions, and source)
- Source value
- 48.1
- Score
- 48.1
- Metric
- Long, w/ CoT
- Unit
- %
- Category
- Long
- Direction
- Higher is better
- Source type
- official-leaderboard
- Evaluator
- LongBench v2 authors, official leaderboard
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact LongBench v2 row QwQ-32B; context column 128k. The source columns w/o CoT and w/ CoT are recorded separately and only where a numeric cell exists. The source-wide YaRN factor-4 note is retained; the leaderboard does not state a per-row 16K thinking budget for this non-Qwen3 row.
- Context
- LongBench v2 does not publish an immutable inference checkpoint/commit, task denominator, or confidence interval for this row.
EFFIBENCH-X (Show measurement, test conditions, and source)
- Source value
- 31.6
- Score
- 31.6
- Metric
- Six-language average Execution Time (ET)
- Unit
- %
- Benchmark version
- EFFIBENCH-X, six-language average, Table 2, arXiv:2505.13004
- Category
- coding-efficiency
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- EFFIBENCH-X paper, arXiv HTML
- Status
- active
- Retrieved at
- 2026-10-19
- Methodology
- Independent evaluation. EFFIBENCH-X, Table 2 six-programming-language average, Execution Time (ET); AWS i7ie.metal-48xl, Docker-isolated code execution, 10-second timeout and 1024 MiB memory limit.
- Context
- The paper does not publish a per-cell sample count, confidence interval, exact weight revision, or generation settings. Percentages are the paper's reported metrics, not a score computed for this record.
EFFIBENCH-X (Show measurement, test conditions, and source)
- Source value
- 35.51
- Score
- 35.51
- Metric
- Six-language average Memory Peak (MP)
- Unit
- %
- Benchmark version
- EFFIBENCH-X, six-language average, Table 2, arXiv:2505.13004
- Category
- coding-efficiency
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- EFFIBENCH-X paper, arXiv HTML
- Status
- active
- Retrieved at
- 2026-10-19
- Methodology
- Independent evaluation. EFFIBENCH-X, Table 2 six-programming-language average, Memory Peak (MP); AWS i7ie.metal-48xl, Docker-isolated code execution, 10-second timeout and 1024 MiB memory limit.
- Context
- The paper does not publish a per-cell sample count, confidence interval, exact weight revision, or generation settings. Percentages are the paper's reported metrics, not a score computed for this record.
EFFIBENCH-X (Show measurement, test conditions, and source)
- Source value
- 31.38
- Score
- 31.38
- Metric
- Six-language average Memory Integral (MI)
- Unit
- %
- Benchmark version
- EFFIBENCH-X, six-language average, Table 2, arXiv:2505.13004
- Category
- coding-efficiency
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- EFFIBENCH-X paper, arXiv HTML
- Status
- active
- Retrieved at
- 2026-10-19
- Methodology
- Independent evaluation. EFFIBENCH-X, Table 2 six-programming-language average, Memory Integral (MI); AWS i7ie.metal-48xl, Docker-isolated code execution, 10-second timeout and 1024 MiB memory limit.
- Context
- The paper does not publish a per-cell sample count, confidence interval, exact weight revision, or generation settings. Percentages are the paper's reported metrics, not a score computed for this record.
EFFIBENCH-X (Show measurement, test conditions, and source)
- Source value
- 36.78
- Score
- 36.78
- Metric
- Six-language average Pass@1
- Unit
- %
- Benchmark version
- EFFIBENCH-X, six-language average, Table 2, arXiv:2505.13004
- Category
- coding-efficiency
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- EFFIBENCH-X paper, arXiv HTML
- Status
- active
- Retrieved at
- 2026-10-19
- Methodology
- Independent evaluation. EFFIBENCH-X, Table 2 six-programming-language average, Pass@1; AWS i7ie.metal-48xl, Docker-isolated code execution, 10-second timeout and 1024 MiB memory limit.
- Context
- The paper does not publish a per-cell sample count, confidence interval, exact weight revision, or generation settings. Percentages are the paper's reported metrics, not a score computed for this record.
Page 1 of 9
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
The 100 most popular models of this category, 0 to 100, higher is more popular
20 of 101 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 7, 2026
QwQ 32BGPT-5.6 Luna 45Hy3 45Nemotron 3 Super 45Nemotron 3 Nano 45
Method changes in the timeline, a jump at these points is no market event:
- 1November 1, 2026: Value only with search or usage data, 30-day attention
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Similar models
More models of the same kind as QwQ 32B, chosen by model line, developer, size, price and release date.
- Qwen 3 30B A3BAlibaba · same developer
- Qwen 3 32BAlibaba · same developer
- Qwen 2.5 Coder 32BAlibaba · same developer
- Qwen2.5-32B-InstructAlibaba · same developer
- Qwen 3 Coder 30B A3BAlibaba · same developer
- Qwen 3 VL 30B A3B ThinkingAlibaba · same developer
- Qwen 3 30B A3B Thinking 2507Alibaba · same developer
- Qwen 3 VL 30B A3BAlibaba · same developer
All specifications
Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
Frequently asked questions about QwQ 32B
What is the knowledge cutoff of QwQ 32B?
April 2024, according to Alibaba. That is 11 months before release.
Stated by Alibaba
AprJunAugOctDecFebKnowledge cutoffApril 2024ReleasedMar 5, 2025
How large is the context window of QwQ 32B?
131,072 tokens. That is how much text QwQ 32B reads at once, about 200 A4 pages.
Page count as a rule of thumb from Artificial Analysis
GLM-5.1200K
DeepSeek-V3 0324164K
QwQ 32BRank 94 of 103131K
GLM-4.5131K
GPT OSS 120B131K
DeepSeek-V3.2128K
DeepSeek-R1128K
Can I run QwQ 32B locally?
Yes, the weights are open.
Stated by Alibaba
- Open weights
- yes
- License
- Apache 2.0
- Parameters
- 32B
- active per token
- n/a
Sources and data date
Show all 30 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
| Evaluator | Test series | Composite indices | Recorded active measurements |
|---|---|---|---|
| Artificial Analysis | 1 | 0 | 1 |
| LMArena | 0 | 0 | 1 |
Without vendor claims and historical or disputed results; those stay labelled in the single values. 1 measurements are not yet assigned to a test series.