Alibaba Cloud / Qwen Team model product
Qwen3.5 35B A3B
5-35B-A3B is a multimodal Mixture-of-Experts model with 35 billion total parameters and 3 billion activated parameters.
Updated Aug 12, 2026. Default version: Qwen3.5-35B-A3B
LLMBoard score52.0Qwen3.5-35B-A3B
Coverage100%81 benchmark families
Context window262.1KTokens
Official input price$0.25Alibaba API
Qwen3.5 35B A3B Specifications
Technical details for the model's default version.
- Version
- Qwen3.5-35B-A3B
- Released
- Feb 24, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 35B
- Context window
- 262.1K
- Max output
- 65.5K
- Inputs
- audio, image, text, video
- Outputs
- text
- Open weights
- Yes
- License
- Apache 2.0
Qwen3.5 35B A3B Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Qwen3.5 35B A3B Benchmark Results
Benchmark scores for Qwen3.5-35B-A3B.
| AndroidWorld_SR | 71.1% | 01 | 8 | 100.0% | C | |
| Hypersim | 0.131 points | 01 | 4 | 100.0% | C | |
| VLMsAreBlind | 97.0% | 01 | 4 | 100.0% | C | |
| CountBench | 0.978 points | 02 | 6 | 80.0% | C | |
| BrowseComp-zh | 69.5% | 03 | 13 | 83.3% | C | |
| FullStackBench en | 58.1% | 03 | 3 | 0.0% | C | |
| FullStackBench zh | 55.0% | 03 | 3 | 0.0% | C | |
| Hallusion Bench | 67.9% | 03 | 16 | 86.7% | C | |
| MMLongBench-Doc | 0.595 points | 03 | 5 | 50.0% | C | |
| MMStar | 81.9% | 03 | 22 | 90.5% | C | |
| MVBench | 74.8% | 03 | 17 | 87.5% | C | |
| Nuscene | 14.6% | 03 | 3 | 0.0% | C | |
| PMC-VQA | 62.0% | 03 | 3 | 0.0% | C | |
| SlakeVQA | 78.7% | 03 | 4 | 33.3% | C | |
| TIR-Bench | 55.5% | 03 | 4 | 33.3% | C | |
| VideoMME w/o sub. | 82.5% | 03 | 10 | 77.8% | C | |
| LingoQA | 79.2% | 04 | 4 | 0.0% | C | |
| MedXpertQA | 61.4% | 04 | 12 | 72.7% | C | |
| MMVU | 72.3% | 04 | 4 | 0.0% | C | |
| OCRBench | 91.0% | 04 | 22 | 85.7% | C | |
| SUNRGBD | 0.334 points | 04 | 4 | 0.0% | C | |
| ZEROBench-Sub | 0.341 points | 04 | 5 | 25.0% | C | |
| CodeForces | 82.2% | 05 | 16 | 73.3% | C | |
| DynaMath | 85.0% | 05 | 7 | 33.3% | C | |
| MMBench-V1.1 | 91.5% | 05 | 18 | 76.5% | C | |
| OJBench | 36.0% | 05 | 9 | 50.0% | C | |
| BFCL-V4 | 67.3% | 06 | 13 | 58.3% | C | |
| DeepPlanning | 22.8% | 06 | 9 | 37.5% | C | |
| MathVista-Mini | 86.2% | 06 | 23 | 77.3% | C | |
| OmniDocBench 1.5 | 89.3% | 06 | 17 | 68.8% | C | |
| RefSpatialBench | 0.635 points | 06 | 6 | 0.0% | C | |
| Seal-0 | 41.4% | 06 | 6 | 0.0% | C | |
| V* | 92.7% | 06 | 7 | 16.7% | C | |
| VideoMME w sub. | 86.6% | 06 | 10 | 44.4% | C | |
| EmbSpatialBench | 0.831 points | 07 | 8 | 14.3% | C | |
| ERQA | 64.8% | 07 | 23 | 72.7% | C | |
| MAXIFE | 86.6% | 07 | 11 | 40.0% | C | |
| MLVU | 85.6% | 07 | 10 | 33.3% | C | |
| NOVA-63 | 57.1% | 07 | 11 | 40.0% | C | |
| PolyMATH | 64.4% | 07 | 23 | 72.7% | C | |
| RefCOCO-avg | 0.892 points | 07 | 7 | 0.0% | C | |
| C-Eval | 90.2% | 08 | 18 | 58.8% | C | |
| HMMT25 | 89.2% | 08 | 25 | 70.8% | C | |
| LVBench | 71.4% | 08 | 24 | 69.6% | C | |
| RealWorldQA | 84.1% | 08 | 26 | 72.0% | C | |
| t2-bench | 81.2% | 08 | 23 | 68.2% | C | |
| VITA-Bench | 31.9% | 08 | 10 | 22.2% | C | |
| ZEROBench | 0.08 points | 08 | 9 | 12.5% | C | |
| AI2D | 92.6% | 09 | 32 | 74.2% | C | |
| BabyVision | 38.4% | 09 | 9 | 0.0% | C | |
| CC-OCR | 80.7% | 09 | 18 | 52.9% | C | |
| Global PIQA | 86.6% | 09 | 13 | 33.3% | C | |
| MMLU-ProX | 81.0% | 09 | 32 | 74.2% | C | |
| WideSearch | 57.1% | 09 | 9 | 0.0% | C | |
| IFEval | 91.9% | 11 | 65 | 84.4% | C | |
| Include | 79.7% | 11 | 31 | 66.7% | C | |
| MathVision | 83.9% | 11 | 32 | 67.7% | C | |
| MMLU-Redux | 93.3% | 11 | 48 | 78.7% | C | |
| Multi-Challenge | 60.0% | 11 | 29 | 64.3% | C | |
| ODinW | 42.6% | 11 | 16 | 33.3% | C | |
| ScreenSpot Pro | 68.6% | 11 | 25 | 58.3% | C | |
| SimpleVQA | 0.583 points | 11 | 13 | 16.7% | C | |
| WMT24++ | 76.3% | 11 | 23 | 54.5% | C | |
| AA-LCR | 58.5% | 12 | 16 | 26.7% | C | |
| LongBench v2 | 59.0% | 12 | 17 | 31.3% | C | |
| MMMU | 81.4% | 13 | 63 | 80.7% | C | |
| MMLU-Pro | 85.3% | 14 | 129 | 89.8% | C | |
| SuperGPQA | 63.4% | 14 | 34 | 60.6% | C | |
| VideoMMMU | 80.4% | 16 | 26 | 40.0% | C | |
| IFBench | 70.2% | 17 | 29 | 42.9% | C | |
| HMMT 2025 | 89.0% | 19 | 33 | 43.8% | C | |
| OSWorld-Verified | 54.5% | 22 | 23 | 4.5% | C | |
| LiveCodeBench v6 | 74.6% | 24 | 53 | 55.8% | C | |
| Humanity's Last Exam | 47.4% | 25 | 93 | 73.9% | C | |
| CharXiv-R | 77.5% | 27 | 48 | 44.7% | C | |
| MMMLU | 85.2% | 32 | 49 | 35.4% | C | |
| MMMU-Pro | 75.1% | 33 | 66 | 50.8% | C | |
| BrowseComp | 61.0% | 35 | 58 | 40.4% | C | |
| Terminal-Bench 2.0 | 40.5% | 45 | 49 | 8.3% | C | |
| GPQA | 84.2% | 61 | 234 | 74.3% | C | |
| SWE-Bench Verified | 69.2% | 64 | 105 | 39.4% | C |
Qwen3.5 35B A3B Arena Results
Preference and agent-evaluation results for the default version.
| webdev | overall | 98 | 1249.6 | 1,540 | N/A | |
| text factuality | overall | 124 | 1408.7 | 28,947 | N/A | |
| text | overall | 136 | 1395.6 | 29,144 | N/A | |
| text style control | overall | 143 | 1395.4 | 29,144 | N/A |
Qwen3.5 35B A3B Runtime Performance
Provider-specific output speed and catalog latency for Qwen3.5-35B-A3B. Runtime does not affect the capability score.
No runtime data
Browse runtime rankingsNo provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Qwen3.5 35B A3B Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.25 input, $2 output per 1M
- Official provider
- Alibaba
- Lowest third-party
- From $0.057 input, $0.459 output per 1M via EmpirioLabs AI
- Tracked offerings
- 14
| EmpirioLabs AI | qwen3-5-35b-a3b | global | $0.057 | $0.459 | 256K | |
| OrcaRouter | qwen/qwen3.5-35b-a3b | global | $0.057 | $0.459 | 262.1K | |
| Merge Gateway | qwen/qwen3.5-35b-a3b | global | $0.057 | $0.459 | 131.1K | |
| OpenRouter | qwen/qwen3.5-35b-a3b | global | $0.14 | $1 | 262.1K | |
| Deep Infra | Qwen/Qwen3.5-35B-A3B | global | $0.14 | $1 | 262.1K | |
| Kilo Gateway | qwen/qwen3.5-35b-a3b | global | $0.1625 | $1.3 | 262.1K | |
| NanoGPT | qwen3.5-35b-a3b | global | $0.225 | $1.8 | 260.1K | |
| SiliconFlow | Qwen/Qwen3.5-35B-A3B | global | $0.24 | $1.8 | 262.1K | |
| Weights & Biases | Qwen/Qwen3.5-35B-A3B | global | $0.25 | $1.25 | 262.1K | |
| Mixlayer | qwen/qwen3.5-35b-a3b | global | $0.25 | $1.3 | 262.1K | |
| Alibaba | qwen3.5-35b-a3b | global | $0.25 | $2 | 262.1K | |
| NovitaAI | qwen/qwen3.5-35b-a3b | global | $0.25 | $2 | 262.1K | |
| Hugging Face | Qwen/Qwen3.5-35B-A3B | global | $0.25 | $2 | 262.1K | |
| Venice AI | qwen3-5-35b-a3b | global | $0.3125 | $1.25 | 256K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Qwen3.5 35B A3B Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| Qwen3.5-35B-A3B | 52.0 | 35B | 262.1K | 65.5K | Yes | Apache 2.0 |
What is Qwen3.5 35B A3B?
Key information about Qwen3.5 35B A3B and its available data.
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.
Data as of 2026-08-11.
Qwen3.5 35B A3B vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Qwen3.5 35B A3B
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
AC
Qwen3.6 35B A3B
Alibaba Cloud / Qwen Team
AC
Qwen3 235B A22B Thinking
Alibaba Cloud / Qwen Team
AC
Qwen3.5 27B
Alibaba Cloud / Qwen Team
AC
Qwen3 VL 235B A22B Thinking
Alibaba Cloud / Qwen Team
AC
Qwen3.5 122B A10B
Alibaba Cloud / Qwen Team
AC
Qwen3.6 27B
Alibaba Cloud / Qwen Team
FAQ
Common questions about Qwen3.5 35B A3B.
When was Qwen3.5 35B A3B released?
Qwen3.5 35B A3B's default version was released on Feb 24, 2026.
How much does Qwen3.5 35B A3B cost?
Qwen3.5 35B A3B's official API price is $0.25 per million input tokens and $2 per million output tokens via Alibaba. The lowest tracked third-party offer starts at $0.057 input and $0.459 output via EmpirioLabs AI.
Who created Qwen3.5 35B A3B?
Qwen3.5 35B A3B was created by Alibaba Cloud / Qwen Team.
What is the context window for Qwen3.5 35B A3B?
The default version has a 262.1K token context window.
Is Qwen3.5 35B A3B open weight?
Yes. The default version is marked as open weight under Apache 2.0.
How many API providers offer Qwen3.5 35B A3B?
14 provider offerings are linked to the default version.
What models should I compare Qwen3.5 35B A3B with?
Nearby ranked alternatives include o3, Nova 2 Pro, Claude Opus 4.1.