Skip to content
aiqbench

Alibaba · model passport

QwQ 32b Preview

qwq-32b-preview
  • Open weights
  • Apache 2.0 licence
  • 32.8B params
  • Released 2024-11-27(curated)
7/27axes covered
Overall · LMArena
1,154#353 of 396
Best percentile
44.3LiveCodeBench (Artificial Analysis)
Cohort standing
29.7median 41.3

What the data says

Benchmarked on 7 of 27 tracked axes. Every figure below is the model's own published result on that board.

Weakest axes

  • 9 pct — Coding (1,174)
  • 11.1 pct — Overall (1,154)
  • 11.9 pct — Hard prompts (1,168)

Uneven: 35.3 percentile points between its best and worst axis — 44.3 pct on LiveCodeBench (Artificial Analysis) versus 9 pct on Coding.

  • LiveCodeBench (Artificial Analysis)44.3 pct33.7%Verify →
  • MMLU-Pro (Artificial Analysis)31.1 pct64.8%Verify →
  • LiveBench27.5 pct40.2%Verify →
  • Instruction following12.6 pct1,145Verify →
  • Hard prompts11.9 pct1,168Verify →
  • Overall11.1 pct1,154Verify →
  • Coding9 pct1,174Verify →
  • WebDev ArenaNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • GPQA DiamondNot benchmarked
  • AIME (math)Not benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • IFBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

RepositoryQwen/QwQ-32B-Preview
Parameters32.8B
ArchitectureQwen2ForCausalLM
LicenceApache 2.0

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Qwen/QwQ-32B-Preview"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • qwq-32b-preview stands at 29.7 (raw mean 21.1) — measured on 7 of 27 axes.

    standingtemplateVerify →

  • qwq-32b-preview stands at 29.7 versus a cohort median of 41.3 across 136 models released within 180 days — below the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →