LLMs TigerLive LLM benchmarks

Qwen3-32B: benchmark scores, price and ranking

Qwen3-32B from Alibaba ranks #475 of 743 LLMs on the LLMs Tiger aggregate benchmark score (46.9), based on 10 benchmark results captured up to 2026-10-13.

Organisation
Alibaba
Type
Open-weight / local model
Open weights
Yes
Released
2025-04-29 (source: curated)
Cheapest API price
$0.080 in / $0.230 out per million tokens (via ovhcloud, observed 2026-10-13T19:47Z)
Aggregate rank
#475 of 743 (score 46.9)

Relative to other models, Qwen3-32B is strongest on Fiction.liveBench (#10 of 47) and weakest on SciCode (#82 of 106).

Across the 10 benchmarks where at least ten models are scored, Qwen3-32B is in the top 10% on 0 and in the top half on 3.

Its cheapest listed input price, $0.080 per million tokens, is lower than that of 85% of the 427 models with a listed price.

Qwen3-32B benchmark results

Qwen3-32B benchmark scores, ranked against all models measured on each benchmark
BenchmarkScoreRankSourceCaptured
Fiction.liveBench74.2%#10 of 47Epoch AI Benchmarking Hub2026-10-13
Epoch ECI139#58 of 204Epoch AI Benchmarking Hub2026-10-13
OTIS Mock AIME66.9%#84 of 189Epoch AI Benchmarking Hub2026-10-13
Aider Polyglot40.0%#26 of 49Aider polyglot leaderboard2026-10-13
DTBench67.5%#80 of 146Epoch AI Benchmarking Hub2026-10-13
GPQA Diamond65.7%#156 of 275Epoch AI Benchmarking Hub2026-10-13
Chess Puzzles5.0%#80 of 122Epoch AI Benchmarking Hub2026-10-13
CritPt0.3%#82 of 115Epoch AI Benchmarking Hub2026-10-13
LMCA17.3%#79 of 108Epoch AI Benchmarking Hub2026-10-13
SciCode35.4%#82 of 106Epoch AI Benchmarking Hub2026-10-13

About Qwen3-32B

Qwen3-32B is an open-weight model that can run on your own hardware from Alibaba. It was released on 2025-04-29, according to the developer's announcement. The cheapest API listing we track charges $0.080 per million input tokens and $0.230 per million output tokens, through ovhcloud. We hold 10 benchmark results for it across 4 areas, and none is in the top tenth of its board.

Where Qwen3-32B stands

On our aggregate score, Qwen3-32B is #475 of 743 ranked models, in the lower half. Its score is 46.9; the scale is centred on 50, so it sits 3.1 points below the typical model. That beats 36% of the ranked field. Among the 74 ranked Alibaba models it is #47. The next one up is qwen3.7-max-20260517 at 47.2. The next one down is qwen3-30b-a3b-instruct-2507 at 46.8.

Qwen3-32B, benchmark by benchmark

Fiction.liveBench puts Qwen3-32B at 74.2%, #10 of 47. That beats 79% of the models tested, but 20.2 points short of grok-4-0709 (94.4%). The results either side are Qwen3-235B-A22B-Thinking-2507 (75.0%) and DeepSeek-R1 (69.4%). It covers long-context story comprehension (16k). The board median is 61.1%, so Qwen3-32B is 13.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

On Epoch ECI, Qwen3-32B reaches 139. It ranks #58 of 204, ahead of 72% of the field, and 18 points behind Qwen 3.8 Max at 156. The results either side are Qwen3-235B-A22B-Instruct (Jul 2025) (139) and grok-3 (138). It covers Epoch Capabilities Index: Epoch's own aggregate of ~40 benchmarks. The board median is 127, so Qwen3-32B is 11 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-32B manages 66.9% on OTIS Mock AIME, good for #84 of 189. Claude Fable 5.1 (max) leads at 100.0%, 33.1 points further up, and 56% of models sit below. The results either side are seed-oss-36b-instruct (67.5%) and claude-haiku-4-5-20251001_32K (66.7%). It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so Qwen3-32B is 9.2 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-32B lands below the middle on Aider Polyglot: 40.0%, #26 of 49, with 51% of the field scoring higher. gemini-2.5-pro-preview-06-05 leads at 83.1%; the closest result above is gpt-4.5-preview at 44.9%. It covers edits across 6 languages (225 exercises). The board median is 44.9%, so Qwen3-32B is 4.9 points below it. Source: Aider polyglot leaderboard, captured 2026-10-13.

Qwen3-32B lands below the middle on DTBench: 67.5%, #80 of 146, with 54% of the field scoring higher. Claude Opus 5.5 (max) leads at 98.9%; the closest result above is mistral-medium-2508 at 67.5%. It covers dynamic tool-use benchmark. The board median is 68.8%, so Qwen3-32B is 1.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-32B lands below the middle on GPQA Diamond: 65.7%, #156 of 275, with 56% of the field scoring higher. GPT-6 Astra (max) leads at 95.8%; the closest result above is Qwen3-4B-Thinking-2507 at 65.8%. It covers expert-level science questions. The board median is 71.7%, so Qwen3-32B is 6.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-32B lands below the middle on Chess Puzzles: 5.0%, #80 of 122, with 65% of the field scoring higher. GPT-6 Astra (max) leads at 72.0%; the closest result above is Gemma 4 31B IT at 5.0%. It covers chess tactics puzzles. The board median is 12.5%, so Qwen3-32B is 7.5 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-32B lands below the middle on CritPt: 0.3%, #82 of 115, with 70% of the field scoring higher. GPT-5.6 Sol (max) leads at 32.3%; the closest result above is qwen3-30b-a3b-thinking-2507 at 0.3%. It covers unpublished research-level physics problems. The board median is 2.9%, so Qwen3-32B is 2.6 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-32B lands below the middle on LMCA: 17.3%, #79 of 108, with 72% of the field scoring higher. Claude Opus 5.5 (max) leads at 68.2%; the closest result above is Llama-3.3-70B-Instruct at 17.5%. It covers language-model capability assessment. The board median is 29.2%, so Qwen3-32B is 11.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-32B lands below the middle on SciCode: 35.4%, #82 of 106, with 76% of the field scoring higher. Claude Opus 5.5 (max) leads at 66.9%; the closest result above is DeepSeek-V3 at 35.4%. It covers scientific research code. The board median is 44.3%, so Qwen3-32B is 8.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Strengths and weaknesses

Software engineering: uneven, ahead of 35% of the field on average across 2 benchmarks. Its best board here is Aider Polyglot (#26 of 49) and its weakest is SciCode (#82 of 106). Reasoning, maths and science: uneven, ahead of 43% of the field on average across 3 benchmarks. Its best board here is OTIS Mock AIME (#84 of 189) and its weakest is CritPt (#82 of 115). General knowledge and chat: solid, ahead of 53% of the field on average across 4 benchmarks. Its best board here is Fiction.liveBench (#10 of 47) and its weakest is LMCA (#79 of 108).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Qwen3-32B costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on Llama 4 Scout ($0.100) and Gemma 4 31B IT ($0.100).

Release history

The Alibaba model released before it was qwen3-235b-a22b on 2025-04-29, which scores 47.9. The next one was qwen3-4b on 2025-04-29, at 47.9.

Reading these numbers

The 10 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (9), Aider polyglot leaderboard (1). 1 of them come from an official leaderboard and 9 from datasets or third-party runs. The leaderboard also lists 1 other reasoning-effort setting of this model, which redirect here. This page shows the highest.

Qwen3-32B is also listed at 1 other reasoning-effort setting; this page shows the highest-effort row, which is the one the leaderboard keeps by default.

Models ranked nearby

More from Alibaba

Compare it on the live leaderboard → · How the aggregate score works