Qwen3-235B-A22B-Thinking-2507: benchmark scores, price and ranking
Qwen3-235B-A22B-Thinking-2507 from Alibaba ranks #243 of 743 LLMs on the LLMs Tiger aggregate benchmark score (49.7), based on 11 benchmark results captured up to 2026-10-13.
- Organisation
- Alibaba
- Type
- Open-weight / local model
- Open weights
- Yes
- Cheapest API price
- $0.300 in / $2.90 out per million tokens (via deepinfra, observed 2026-10-13T19:47Z)
- Aggregate rank
- #243 of 743 (score 49.7)
Relative to other models, Qwen3-235B-A22B-Thinking-2507 is strongest on Fiction.liveBench (#9 of 47) and weakest on CritPt (#112 of 115).
Across the 11 benchmarks where at least ten models are scored, Qwen3-235B-A22B-Thinking-2507 is in the top 10% on 0 and in the top half on 5.
Its cheapest listed input price, $0.300 per million tokens, is lower than that of 58% of the 427 models with a listed price.
Qwen3-235B-A22B-Thinking-2507 benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| Fiction.liveBench | 75.0% | #9 of 47 | Epoch AI Benchmarking Hub | 2026-10-13 |
| OTIS Mock AIME | 86.7% | #53 of 189 | Epoch AI Benchmarking Hub | 2026-10-13 |
| DTBench | 80.3% | #53 of 146 | Epoch AI Benchmarking Hub | 2026-10-13 |
| GPQA Diamond | 80.1% | #105 of 275 | Epoch AI Benchmarking Hub | 2026-10-13 |
| LMCA | 29.3% | #54 of 108 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WeirdML | 41.0% | #56 of 109 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Chess Puzzles | 12.0% | #65 of 122 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SimpleQA Verified | 40.4% | #37 of 69 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 42.4% | #57 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Mystery Games | 9.0% | #46 of 59 | Epoch AI Benchmarking Hub | 2026-10-13 |
| CritPt | 0.0% | #112 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
About Qwen3-235B-A22B-Thinking-2507
Qwen3-235B-A22B-Thinking-2507 is an open-weight model that can run on your own hardware from Alibaba. The cheapest API listing we track charges $0.300 per million input tokens and $2.90 per million output tokens, through deepinfra. We hold 11 benchmark results for it across 4 areas, and none is in the top tenth of its board.
Where Qwen3-235B-A22B-Thinking-2507 stands
On our aggregate score, Qwen3-235B-A22B-Thinking-2507 is #243 of 743 ranked models, around the middle of the table. Its score is 49.7; the scale is centred on 50, so it sits 0.3 points below the typical model. That beats 67% of the ranked field. Among the 74 ranked Alibaba models it is #19. The next one up is QwQ-32B at 49.7. The next one down is ollama/qwen2.5:32b-instruct-q8_0 at 49.6.
Qwen3-235B-A22B-Thinking-2507, benchmark by benchmark
Fiction.liveBench puts Qwen3-235B-A22B-Thinking-2507 at 75.0%, #9 of 47. That beats 81% of the models tested, but 19.4 points short of grok-4-0709 (94.4%). The results either side are DeepSeek-R1 (May 2025) (75.0%) and Qwen3-32B (74.2%). It covers long-context story comprehension (16k). The board median is 61.1%, so Qwen3-235B-A22B-Thinking-2507 is 13.9 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On OTIS Mock AIME, Qwen3-235B-A22B-Thinking-2507 reaches 86.7%. It ranks #53 of 189, ahead of 72% of the field, and 13.3 points behind Claude Fable 5.1 (max) at 100.0%. The results either side are nemotron-3-ultra (86.7%) and Qwen3.5 Plus (86.7%). It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so Qwen3-235B-A22B-Thinking-2507 is 28.9 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Qwen3-235B-A22B-Thinking-2507 manages 80.3% on DTBench, good for #53 of 146. Claude Opus 5.5 (max) leads at 98.9%, 18.7 points further up, and 64% of models sit below. The results either side are GPT-5.4 nano (xhigh) (80.3%) and Claude Opus 4.1 (unknown thinking) (80.0%). It covers dynamic tool-use benchmark. The board median is 68.8%, so Qwen3-235B-A22B-Thinking-2507 is 11.5 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
GPQA Diamond puts Qwen3-235B-A22B-Thinking-2507 at 80.1%, #105 of 275. That beats 62% of the models tested, but 15.7 points short of GPT-6 Astra (max) (95.8%). The results either side are Claude Sonnet 5 (max) (80.3%) and Claude 3.7 Sonnet (64k thinking) (79.7%). It covers expert-level science questions. The board median is 71.7%, so Qwen3-235B-A22B-Thinking-2507 is 8.3 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On LMCA, Qwen3-235B-A22B-Thinking-2507 reaches 29.3%. It ranks #54 of 108, ahead of 50% of the field, and 38.9 points behind Claude Opus 5.5 (max) at 68.2%. The results either side are qwen3.5-35b-a3b (29.5%) and Qwen3.5 Flash (29.1%). It covers language-model capability assessment. The board median is 29.2%, so Qwen3-235B-A22B-Thinking-2507 is 0.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Qwen3-235B-A22B-Thinking-2507 lands below the middle on WeirdML: 41.0%, #56 of 109, with 50% of the field scoring higher. GPT-6 Astra (max) leads at 93.3%; the closest result above is Qwen3-Coder-480B-A35B-Instruct at 41.2%. It covers unusual machine-learning tasks. The board median is 41.2%, so Qwen3-235B-A22B-Thinking-2507 is 0.1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Qwen3-235B-A22B-Thinking-2507 lands below the middle on SimpleQA Verified: 40.4%, #37 of 69, with 52% of the field scoring higher. gemini-4-argon-high leads at 76.0%; the closest result above is GPT-5.6 Luna (max) at 41.0%. It covers short factual questions. The board median is 41.0%, so Qwen3-235B-A22B-Thinking-2507 is 0.6 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Qwen3-235B-A22B-Thinking-2507 lands below the middle on SciCode: 42.4%, #57 of 106, with 53% of the field scoring higher. Claude Opus 5.5 (max) leads at 66.9%; the closest result above is mimo-v2.5 at 43.1%. It covers scientific research code. The board median is 44.3%, so Qwen3-235B-A22B-Thinking-2507 is 2.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Qwen3-235B-A22B-Thinking-2507 lands below the middle on Mystery Games: 9.0%, #46 of 59, with 76% of the field scoring higher. GPT-6 Astra (max) leads at 84.0%; the closest result above is Qwen 3.6 Plus (2026-04-02) at 12.0%. It covers deductive reasoning puzzles. The board median is 23.0%, so Qwen3-235B-A22B-Thinking-2507 is 14.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
CritPt is a weak spot. Qwen3-235B-A22B-Thinking-2507 scores 0.0%, #112 of 115, behind 97% of the models measured. The leader, GPT-5.6 Sol (max), reaches 32.3%. It covers unpublished research-level physics problems. The board median is 2.9%, so Qwen3-235B-A22B-Thinking-2507 is 2.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
The full table above also lists the other 1 result, including boards with fewer than ten models, where a rank says little.
Strengths and weaknesses
Reasoning, maths and science: uneven, ahead of 45% of the field on average across 3 benchmarks. Its best board here is OTIS Mock AIME (#53 of 189) and its weakest is CritPt (#112 of 115). General knowledge and chat: uneven, ahead of 49% of the field on average across 6 benchmarks. Its best board here is Fiction.liveBench (#9 of 47) and its weakest is Mystery Games (#46 of 59).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so Qwen3-235B-A22B-Thinking-2507 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), Qwen3.7 Flash ($0.030, 50.5). Similar scores cost more on qwen-2.5-72b-instruct ($0.380) and Devstral (2512) ($0.400).
Reading these numbers
The 11 scores on this page come from 1 source: Epoch AI Benchmarking Hub (11). No official leaderboard supplied any of them, so treat them as third-party measurements.
Models ranked nearby
- #240 Kimi K2 Thinking (Together) (49.7)
- #241 Frogboss 32B 2510 (49.7)
- #242 claude-sonnet-4-20250514_16K (49.7)
- #244 Inkling Small (xhigh) (49.7)
- #245 GLM-4.5-FP8 (49.6)
- #246 GLM-5.1 (49.6)
More from Alibaba
- Qwen 3.6 Max (Preview) (#25, 54.1)
- Qwen3.8 Max (xhigh) (#37, 53.2)
- Qwen3.7 Max (#71, 52.3)
- ollama/qwen2.5-coder:32b (#93, 51.9)
- Qwen3 235B A22B diff, no think, Alibaba API (#122, 51.5)
- Qwen2.5-Coder-14B-Instruct (#135, 51.3)
Compare it on the live leaderboard → · How the aggregate score works