Rankings / Alibaba
Qwen3 235B A22b
released 2025-04-29
BenchAtlas Index
as of 2026-09-11
base configuration
rank #322
5 families · 5 categories · medium
base configuration
rank #330
7 families · 4 categories · medium
thinking
rank #346
10 families · 5 categories · high
Benchmark evidence
65 results
agentic coding
| τ²-Bench | no reasoning | 27.2 % | independent | Artificial Analysis |
| τ²-Bench | thinking | 24.0 % | independent | Artificial Analysis |
coding
| LiveCodeBench v6 problems=454 · window_start=2024-08-01 | 65.9 % | independent | LiveCodeBench Leaderboard | |
| LiveCodeBench v6 implementation=artificial-analysis | thinking | 62.2 % | independent | Artificial Analysis |
| LiveCodeBench v6 implementation=artificial-analysis | no reasoning | 34.3 % | independent | Artificial Analysis |
| SciCode | thinking | 39.9 % | independent | Artificial Analysis |
| SciCode | no reasoning | 29.9 % | independent | Artificial Analysis |
external indices
| Intelligence Index v4.1 | thinking | 9.5 points | independent | Artificial Analysis |
| Intelligence Index v4.1 | no reasoning | 8.3 points | independent | Artificial Analysis |
| AA Math Index | thinking | 82.0 points | independent | Artificial Analysis |
| AA Math Index | no reasoning | 23.7 points | independent | Artificial Analysis |
| ECI | 139.4 points | independent | Epoch AI Benchmarking Hub |
factuality
| MASK contamination=Potential contamination warning: This model was evaluated after the public release of MASK, allowing model builder access to the prompts and solutions. | 56.4 % 51.4–61.4 | independent | Scale Labs |
human preference
| Coding (style control)older version arena=text · category=coding · style_control=true | 1432.3 1423.3–1441.4 | community | LMArena Leaderboard Dataset | |
| Industry Software And It Services (style control)older version arena=text · category=industry_software_and_it_services · style_control=true | 1416.1 1409.1–1423.2 | community | LMArena Leaderboard Dataset | |
| Hard Prompts English (style control)older version arena=text · category=hard_prompts_english · style_control=true | 1401.5 1393.2–1409.8 | community | LMArena Leaderboard Dataset | |
| Industry Mathematical (style control)older version arena=text · category=industry_mathematical · style_control=true | 1398.5 1383.5–1413.5 | community | LMArena Leaderboard Dataset | |
| Chinese (style control)older version arena=text · category=chinese · style_control=true | 1391.8 1376.9–1406.7 | community | LMArena Leaderboard Dataset | |
| Math (style control)older version arena=text · category=math · style_control=true | 1391.7 1377.8–1405.7 | community | LMArena Leaderboard Dataset | |
| Hard Prompts (style control)older version arena=text · category=hard_prompts · style_control=true | 1391.2 1384.9–1397.5 | community | LMArena Leaderboard Dataset | |
| English (style control)older version arena=text · category=english · style_control=true | 1387.2 1381.1–1393.2 | community | LMArena Leaderboard Dataset | |
| Expert (style control)older version arena=text · category=expert · style_control=true | 1385.4 1369.5–1401.4 | community | LMArena Leaderboard Dataset | |
| German (style control)older version arena=text · category=german · style_control=true | 1385.1 1363.4–1406.8 | community | LMArena Leaderboard Dataset | |
| Industry Life And Physical And Social Science (style control)older version arena=text · category=industry_life_and_physical_and_social_science · style_control=true | 1384.0 1375.0–1393.0 | community | LMArena Leaderboard Dataset | |
| Industry Medicine And Healthcare (style control)older version arena=text · category=industry_medicine_and_healthcare · style_control=true | 1382.1 1366.8–1397.4 | community | LMArena Leaderboard Dataset | |
| Longer Query (style control)older version arena=text · category=longer_query · style_control=true | 1378.4 1369.6–1387.1 | community | LMArena Leaderboard Dataset | |
| Overall (style control) arena=text · category=overall · style_control=true | 1374.9 1370.2–1379.7 | community | LMArena Leaderboard Dataset | |
| Multi Turn (style control)older version arena=text · category=multi_turn · style_control=true | 1370.4 1361.1–1379.6 | community | LMArena Leaderboard Dataset | |
| Industry Business And Management And Financial Operations (style control)older version arena=text · category=industry_business_and_management_and_financial_operations · style_control=true | 1369.5 1360.1–1379.0 | community | LMArena Leaderboard Dataset | |
| Industry Legal And Government (style control)older version arena=text · category=industry_legal_and_government · style_control=true | 1365.8 1351.8–1379.8 | community | LMArena Leaderboard Dataset | |
| Polish (style control)older version arena=text · category=polish · style_control=true | 1361.0 1346.9–1375.1 | community | LMArena Leaderboard Dataset | |
| Russian (style control)older version arena=text · category=russian · style_control=true | 1357.6 1345.7–1369.6 | community | LMArena Leaderboard Dataset | |
| Instruction Following (style control)older version arena=text · category=instruction_following · style_control=true | 1356.5 1348.8–1364.3 | community | LMArena Leaderboard Dataset | |
| Non English (style control)older version arena=text · category=non_english · style_control=true | 1356.0 1350.2–1361.7 | community | LMArena Leaderboard Dataset | |
| Exclude Ties (style control)older version arena=text · category=exclude_ties · style_control=true | 1345.6 1339.0–1352.2 | community | LMArena Leaderboard Dataset | |
| Industry Writing And Literature And Language (style control)older version arena=text · category=industry_writing_and_literature_and_language · style_control=true | 1344.9 1337.0–1352.8 | community | LMArena Leaderboard Dataset | |
| Creative Writing (style control)older version arena=text · category=creative_writing · style_control=true | 1323.3 1313.2–1333.4 | community | LMArena Leaderboard Dataset | |
| Industry Entertainment And Sports And Media (style control)older version arena=text · category=industry_entertainment_and_sports_and_media · style_control=true | 1322.9 1314.3–1331.6 | community | LMArena Leaderboard Dataset | |
| Japanese (style control)older version arena=text · category=japanese · style_control=true | 1314.9 1291.5–1338.3 | community | LMArena Leaderboard Dataset |
knowledge science
| GPQA Diamond | 70.7 % | independent | Epoch AI Benchmarking Hub 2025-06-03 | |
| GPQA Diamond implementation=vals-ai | 70.2 % | independent | Vals AI | |
| GPQA Diamond implementation=artificial-analysis | thinking | 70.0 % | independent | Artificial Analysis |
| GPQA Diamond implementation=artificial-analysis | no reasoning | 61.3 % | independent | Artificial Analysis |
| Humanity's Last Exam implementation=artificial-analysis | thinking | 11.0 % | independent | Artificial Analysis |
| Humanity's Last Exam implementation=artificial-analysis | no reasoning | 4.2 % | independent | Artificial Analysis |
| MMLU-Pro implementation=artificial-analysis | thinking | 82.8 % | independent | Artificial Analysis |
| MMLU-Pro implementation=vals-ai | 81.2 % | independent | Vals AI | |
| MMLU-Pro implementation=artificial-analysis | no reasoning | 76.2 % | independent | Artificial Analysis |
long context instruction
| AA-LCR | thinking | 0.0 % | independent | Artificial Analysis |
| AA-LCR | no reasoning | 0.0 % | independent | Artificial Analysis |
| IFBench | thinking | 38.7 % | independent | Artificial Analysis |
| IFBench | no reasoning | 36.6 % | independent | Artificial Analysis |
| MultiChallenge | 41.2 % 39.4–43.0 | independent | Scale Labs |
professional
| LegalBench | 80.2 % | independent | Vals AI | |
| MedQA | 90.6 % | independent | Vals AI | |
| TaxEval | 70.6 % | independent | Vals AI |
reasoning math
| AIME implementation=artificial-analysis | thinking | 84.0 % | independent | Artificial Analysis |
| AIME implementation=vals-ai | 84.0 % | independent | Vals AI | |
| AIME year=2025 · implementation=artificial-analysis | thinking | 82.0 % | independent | Artificial Analysis |
| AIME implementation=artificial-analysis | no reasoning | 32.7 % | independent | Artificial Analysis |
| AIME year=2025 · implementation=artificial-analysis | no reasoning | 23.7 % | independent | Artificial Analysis |
| MATH-500 implementation=vals-ai | 94.6 % | independent | Vals AI | |
| MATH-500 implementation=artificial-analysis | thinking | 93.0 % | independent | Artificial Analysis |
| MATH-500 implementation=artificial-analysis | no reasoning | 90.2 % | independent | Artificial Analysis |
| MATH Level 5 | 68.9 % | independent | Epoch AI Benchmarking Hub 2025-06-03 |
Agent + model results
systems, not bare-model scores
| agent + model Artificial Analysis harness + Qwen3 235B A22b | Terminal-Bench Hard | 6.1 % | independent | Artificial Analysis |
| agent + model Artificial Analysis harness + Qwen3 235B A22b | Terminal-Bench Hard | 6.1 % | independent | Artificial Analysis |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.