Alibaba Cloud / Qwen Team model product
Qwen3.5-35B-A3B is a multimodal Mixture-of-Experts model with 35 billion total parameters and 3 billion activated parameters. It combines strong reasoning, coding, agentic, and visual understanding performance with production-friendly efficiency and a native 262K context window.
Updated Aug 10, 2026. Default version: Qwen3.5-35B-A3B
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version Qwen3.5-35B-A3B.
| AndroidWorld_SR | 0.7 | 1 | 8 | 100.0% | C | |
| Hypersim | 0.1 | 1 | 4 | 100.0% | C | |
| VLMsAreBlind | 1.0 | 1 | 4 | 100.0% | C | |
| CountBench | 1.0 | 2 | 6 | 80.0% | C | |
| BrowseComp-zh | 0.7 | 3 | 13 | 83.3% | C | |
| FullStackBench en | 0.6 | 3 | 3 | 0.0% | C | |
| FullStackBench zh | 0.6 | 3 | 3 | 0.0% | C | |
| Hallusion Bench | 0.7 | 3 | 16 | 86.7% | C | |
| MMLongBench-Doc | 0.6 | 3 | 5 | 50.0% | C | |
| MMStar | 0.8 | 3 | 22 | 90.5% | C | |
| MVBench | 0.7 | 3 | 17 | 87.5% | C | |
| Nuscene | 0.1 | 3 | 3 | 0.0% | C | |
| PMC-VQA | 0.6 | 3 | 3 | 0.0% | C | |
| SlakeVQA | 0.8 | 3 | 4 | 33.3% | C | |
| TIR-Bench | 0.6 | 3 | 4 | 33.3% | C | |
| VideoMME w/o sub. | 0.8 | 3 | 10 | 77.8% | C | |
| LingoQA | 0.8 | 4 | 4 | 0.0% | C | |
| MedXpertQA | 0.6 | 4 | 12 | 72.7% | C | |
| MMVU | 0.7 | 4 | 4 | 0.0% | C | |
| OCRBench | 0.9 | 4 | 22 | 85.7% | C | |
| SUNRGBD | 0.3 | 4 | 4 | 0.0% | C | |
| ZEROBench-Sub | 0.3 | 4 | 5 | 25.0% | C | |
| CodeForces | 0.8 | 5 | 16 | 73.3% | C | |
| DynaMath | 0.8 | 5 | 7 | 33.3% | C | |
| MMBench-V1.1 | 0.9 | 5 | 18 | 76.5% | C | |
| OJBench | 0.4 | 5 | 9 | 50.0% | C | |
| BFCL-V4 | 0.7 | 6 | 13 | 58.3% | C | |
| DeepPlanning | 0.2 | 6 | 9 | 37.5% | C | |
| MathVista-Mini | 0.9 | 6 | 23 | 77.3% | C | |
| OmniDocBench 1.5 | 0.9 | 6 | 16 | 66.7% | C | |
| RefSpatialBench | 0.6 | 6 | 6 | 0.0% | C | |
| Seal-0 | 0.4 | 6 | 6 | 0.0% | C | |
| V* | 0.9 | 6 | 7 | 16.7% | C | |
| VideoMME w sub. | 0.9 | 6 | 10 | 44.4% | C | |
| EmbSpatialBench | 0.8 | 7 | 8 | 14.3% | C | |
| ERQA | 0.6 | 7 | 23 | 72.7% | C | |
| MAXIFE | 0.9 | 7 | 11 | 40.0% | C | |
| MLVU | 0.9 | 7 | 10 | 33.3% | C | |
| NOVA-63 | 0.6 | 7 | 11 | 40.0% | C | |
| PolyMATH | 0.6 | 7 | 23 | 72.7% | C | |
| RefCOCO-avg | 0.9 | 7 | 7 | 0.0% | C | |
| C-Eval | 0.9 | 8 | 18 | 58.8% | C | |
| HMMT25 | 0.9 | 8 | 25 | 70.8% | C | |
| LVBench | 0.7 | 8 | 24 | 69.6% | C | |
| RealWorldQA | 0.8 | 8 | 26 | 72.0% | C | |
| t2-bench | 0.8 | 8 | 23 | 68.2% | C | |
| VITA-Bench | 0.3 | 8 | 10 | 22.2% | C | |
| ZEROBench | 0.1 | 8 | 9 | 12.5% | C | |
| AI2D | 0.9 | 9 | 32 | 74.2% | C | |
| BabyVision | 0.4 | 9 | 9 | 0.0% | C | |
| CC-OCR | 0.8 | 9 | 18 | 52.9% | C | |
| Global PIQA | 0.9 | 9 | 13 | 33.3% | C | |
| MMLU-ProX | 0.8 | 9 | 32 | 74.2% | C | |
| WideSearch | 0.6 | 9 | 9 | 0.0% | C | |
| ScreenSpot Pro | 0.7 | 10 | 24 | 60.9% | C | |
| AA-LCR | 0.6 | 11 | 15 | 28.6% | C | |
| IFEval | 0.9 | 11 | 65 | 84.4% | C | |
| Include | 0.8 | 11 | 31 | 66.7% | C | |
| MathVision | 0.8 | 11 | 32 | 67.7% | C | |
| MMLU-Redux | 0.9 | 11 | 48 | 78.7% | C | |
| Multi-Challenge | 0.6 | 11 | 29 | 64.3% | C | |
| ODinW | 0.4 | 11 | 16 | 33.3% | C | |
| SimpleVQA | 0.6 | 11 | 13 | 16.7% | C | |
| WMT24++ | 0.8 | 11 | 23 | 54.5% | C | |
| LongBench v2 | 0.6 | 12 | 17 | 31.3% | C | |
| MMMU | 0.8 | 13 | 63 | 80.7% | C | |
| MMLU-Pro | 0.9 | 14 | 129 | 89.8% | C | |
| SuperGPQA | 0.6 | 14 | 34 | 60.6% | C | |
| IFBench | 0.7 | 16 | 28 | 44.4% | C | |
| VideoMMMU | 0.8 | 16 | 26 | 40.0% | C | |
| HMMT 2025 | 0.9 | 19 | 33 | 43.8% | C | |
| OSWorld-Verified | 0.5 | 21 | 22 | 4.8% | C | |
| LiveCodeBench v6 | 0.7 | 24 | 53 | 55.8% | C | |
| Humanity's Last Exam | 0.5 | 25 | 92 | 73.6% | C | |
| CharXiv-R | 0.8 | 26 | 47 | 45.6% | C | |
| MMMLU | 0.9 | 32 | 49 | 35.4% | C | |
| MMMU-Pro | 0.8 | 33 | 65 | 50.0% | C | |
| BrowseComp | 0.6 | 35 | 58 | 40.4% | C | |
| Terminal-Bench 2.0 | 0.4 | 45 | 49 | 8.3% | C | |
| GPQA | 0.8 | 61 | 233 | 74.1% | C | |
| SWE-Bench Verified | 0.7 | 63 | 104 | 39.8% | C |
Preference and agent-evaluation signals from published Arena datasets.
| text factuality | chinese | 48 | 1485.1 | 1,096 | N/A | |
| text factuality | industry mathematical | 76 | 1413.1 | 1,031 | N/A | |
| text factuality | industry medicine and healthcare | 76 | 1460.4 | 1,449 | N/A | |
| text factuality | math | 77 | 1412.5 | 1,147 | N/A | |
| text | chinese | 82 | 1466.3 | 1,592 | N/A | |
| webdev | webdev-react | 87 | 1235.1 | 1,336 | N/A | |
| text factuality | industry legal and government | 89 | 1431.7 | 1,547 | N/A | |
| webdev | webdev-html | 91 | 1297.5 | 185 | N/A | |
| text factuality | expert | 97 | 1426.4 | 1,773 | N/A | |
| webdev | overall | 97 | 1250.4 | 1,521 | N/A | |
| webdev | webdev | 97 | 1250.4 | 1,521 | N/A | |
| text factuality | russian | 100 | 1399.7 | 2,065 | N/A | |
| text | industry medicine and healthcare | 103 | 1425.8 | 2,148 | N/A | |
| text factuality | industry life and physical and social science | 105 | 1436.4 | 3,417 | N/A | |
| text style control | chinese | 106 | 1460.3 | 1,592 | N/A | |
| text factuality | industry business and management and financial operations | 109 | 1418.4 | 4,116 | N/A | |
| text factuality | creative writing | 111 | 1365.0 | 3,087 | N/A | |
| text factuality | multi turn | 111 | 1405.4 | 3,684 | N/A | |
| text | french | 113 | 1416.4 | 825 | N/A | |
| text factuality | coding | 113 | 1449.2 | 6,129 | N/A | |
| text | japanese | 114 | 1329.9 | 224 | N/A | |
| text | korean | 114 | 1352.4 | 465 | N/A | |
| text factuality | industry entertainment and sports and media | 114 | 1365.8 | 4,044 | N/A | |
| text factuality | instruction following | 114 | 1400.5 | 7,460 | N/A | |
| text factuality | hard prompts english | 115 | 1438.2 | 7,176 | N/A | |
| text factuality | industry writing and literature and language | 115 | 1382.5 | 4,812 | N/A | |
| text factuality | longer query | 115 | 1416.5 | 9,214 | N/A | |
| text factuality | industry software and it services | 116 | 1443.3 | 10,380 | N/A | |
| text | polish | 117 | 1387.4 | 553 | N/A | |
| text style control | french | 118 | 1424.7 | 825 | N/A | |
| text | expert | 122 | 1406.6 | 2,515 | N/A | |
| text | spanish | 122 | 1391.7 | 884 | N/A | |
| text factuality | english | 123 | 1420.6 | 13,386 | N/A | |
| text factuality | overall | 124 | 1408.8 | 28,669 | N/A | |
| text factuality | exclude ties | 124 | 1396.2 | 20,935 | N/A | |
| text | industry business and management and financial operations | 125 | 1393.7 | 5,810 | N/A | |
| text | math | 125 | 1405.0 | 1,746 | N/A | |
| text factuality | hard prompts | 126 | 1425.4 | 18,030 | N/A | |
| text factuality | non english | 126 | 1392.4 | 15,152 | N/A | |
| text | industry mathematical | 127 | 1406.7 | 1,561 | N/A | |
| text style control | industry medicine and healthcare | 128 | 1430.8 | 2,148 | N/A | |
| text style control | korean | 128 | 1343.7 | 465 | N/A | |
| text style control | japanese | 129 | 1321.2 | 224 | N/A | |
| text | industry software and it services | 131 | 1415.7 | 11,371 | N/A | |
| text | instruction following | 131 | 1380.3 | 9,205 | N/A | |
| text | hard prompts english | 133 | 1409.1 | 8,880 | N/A | |
| text | industry life and physical and social science | 133 | 1412.3 | 4,879 | N/A | |
| text style control | expert | 133 | 1419.0 | 2,515 | N/A | |
| text style control | spanish | 133 | 1385.4 | 884 | N/A | |
| text | english | 134 | 1407.9 | 13,573 | N/A | |
| text style control | polish | 134 | 1375.6 | 553 | N/A | |
| text style control | industry business and management and financial operations | 135 | 1402.2 | 5,810 | N/A | |
| text style control | math | 135 | 1400.0 | 1,746 | N/A | |
| text | overall | 136 | 1395.6 | 28,864 | N/A | |
| text | coding | 136 | 1409.4 | 7,896 | N/A | |
| text | exclude ties | 136 | 1374.4 | 21,082 | N/A | |
| text | industry legal and government | 136 | 1398.0 | 2,250 | N/A | |
| text | longer query | 136 | 1391.9 | 10,952 | N/A | |
| text | non english | 136 | 1379.0 | 15,291 | N/A | |
| text | hard prompts | 137 | 1400.6 | 18,160 | N/A | |
| text | russian | 138 | 1378.4 | 3,101 | N/A | |
| text style control | industry mathematical | 138 | 1407.0 | 1,561 | N/A | |
| text style control | instruction following | 138 | 1388.2 | 9,205 | N/A | |
| text | multi turn | 139 | 1390.1 | 5,166 | N/A | |
| text style control | english | 139 | 1411.3 | 13,573 | N/A | |
| text style control | industry life and physical and social science | 139 | 1415.2 | 4,879 | N/A | |
| text style control | industry software and it services | 140 | 1430.8 | 11,371 | N/A | |
| text | german | 142 | 1358.3 | 475 | N/A | |
| text | industry writing and literature and language | 142 | 1366.1 | 6,605 | N/A | |
| text style control | exclude ties | 142 | 1375.3 | 21,082 | N/A | |
| text style control | overall | 143 | 1395.3 | 28,864 | N/A | |
| text style control | hard prompts english | 143 | 1423.8 | 8,880 | N/A | |
| text | creative writing | 144 | 1347.9 | 4,412 | N/A | |
| text | industry entertainment and sports and media | 145 | 1347.5 | 5,598 | N/A | |
| text style control | industry legal and government | 145 | 1399.6 | 2,250 | N/A | |
| text style control | longer query | 145 | 1401.6 | 10,952 | N/A | |
| text style control | multi turn | 145 | 1394.5 | 5,166 | N/A | |
| text style control | non english | 147 | 1375.5 | 15,291 | N/A | |
| text style control | russian | 147 | 1377.7 | 3,101 | N/A | |
| text style control | hard prompts | 149 | 1413.2 | 18,160 | N/A | |
| text style control | coding | 150 | 1434.6 | 7,896 | N/A | |
| text style control | industry writing and literature and language | 153 | 1365.3 | 6,605 | N/A | |
| text style control | german | 159 | 1348.8 | 475 | N/A | |
| text style control | industry entertainment and sports and media | 161 | 1345.0 | 5,598 | N/A | |
| text style control | creative writing | 163 | 1343.4 | 4,412 | N/A |
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| NanoGPT | qwen3.5-35b-a3b | global | $0.225 | $1.8 | 260.1K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| Qwen3.5-35B-A3B | 53.3 | 35B | 262.1K | 65.5K | Yes | Apache 2.0 |
A concise description based on the published model registry.
Qwen3.5-35B-A3B is a multimodal Mixture-of-Experts model with 35 billion total parameters and 3 billion activated parameters. It combines strong reasoning, coding, agentic, and visual understanding performance with production-friendly efficiency and a native 262K context window.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about Qwen3.5 35B A3B.
Qwen3.5 35B A3B's default version was released on Feb 24, 2026.
No official standard PAYG price is currently available for Qwen3.5 35B A3B. The lowest tracked third-party offer starts at $0.225 input and $1.8 output via NanoGPT.
Qwen3.5 35B A3B is published under Alibaba Cloud / Qwen Team in the model registry.
The default version has a 262.1K token context window.
Yes. The default version is marked as open weight under Apache 2.0.
1 published provider offerings are linked to the default version.
Nearby ranked alternatives include MiniMax M2.1, Llama 3.3 70B, Qwen3 Next 80B A3B Thinking.