Trinity Mini
- Context
- 262Ktokens
- Released
- 8mo ago
- Size
- 35Bparameters
Pre-trained Qwen3.5 base checkpoint for fine-tuning and research, without instruction tuning
- Multimodal model
- Open weights · 35B
- Accepts images, video
How it compares
Size class
Medium35BAn open-weight model of 25 to 100 billion parameters, for a workstation or a single server GPU.
Open models of a similar size from other labs
Vs. your stack
Add the models you use to see how this one compares.
Sign inAdditional source evidence
- Source model
- alibaba/qwen3.5-35b-a3b-base
- Source family
- qwen
- Last updated
- 2026-02-24
- Model context
- 262K
- Model output limit
- 66K
- Open weights
- Yes
- File attachments
- Yes
- Reasoning
- No
- Tool calling
- No
- Temperature control
- Yes
Downloadable weights
Complete source record (JSON)
{
"id": "alibaba/qwen3.5-35b-a3b-base",
"name": "Qwen3.5 35B-A3B Base",
"description": "Pre-trained Qwen3.5 base checkpoint for fine-tuning and research, without instruction tuning",
"family": "qwen",
"last_updated": "2026-02-24",
"attachment": true,
"reasoning": false,
"tool_call": false,
"temperature": true,
"modalities": {
"input": [
"text",
"image",
"video"
],
"output": [
"text"
]
},
"limit": {
"context": 262144,
"output": 65536
},
"release_date": "2026-02-24",
"open_weights": true,
"license": "Apache-2.0",
"weights": [
{
"label": "Hugging Face",
"url": "https://huggingface.co/Qwen/Qwen3.5-35B-A3B-Base"
}
]
}Frequently asked questions
Running locally
Qwen3.5 35B-A3B Base needs an estimated 24.9 GB of VRAM at Q4, 43.3 GB at Q8 or 81 GB at FP16. These figures are estimates from the 35B parameter count, covering weights plus about 15% and 0.5 GB for a short context.
No. A 24 GB, 16 GB, 12 GB or 8 GB GPU does not fit Qwen3.5 35B-A3B Base at Q4, Q8 or FP16, based on estimates.
Yes, with enough unified memory. Q4 needs an estimated 24.9 GB, and assuming macOS lets the GPU use about 75% of unified memory, the smallest common Mac memory size that fits Q4 is 36 GB (estimate).
Privacy and data
Yes. Qwen3.5 35B-A3B Base has open weights, so you can self-host it for full control of your data.
Capabilities and specs
The context window is 262K tokens (262,144), and the maximum output is 66K tokens (65,536).
Yes for images and video. Input modalities are text, image and video, while audio and PDFs are not listed as supported inputs.
No. Tool calling (function calling) is not supported.
No. A reasoning or thinking mode is not supported.
Quality and benchmarks
Olmo Hybrid 7B Instruct and Olmo Hybrid 7B Think by Ai2, and MiniMax-M2.5-highspeed by MiniMax. They were matched on model kind and release date, with no comparable price.
Olmo Hybrid 7B InstructQwen3.5 35B-A3B Base vs Olmo Hybrid 7B InstructOlmo Hybrid 7B ThinkQwen3.5 35B-A3B Base vs Olmo Hybrid 7B ThinkMiniMax-M2.5-highspeedQwen3.5 35B-A3B Base vs MiniMax-M2.5-highspeed
Versions and lineage
Both share a context window of 262K tokens (262,144). The difference is the release date, Qwen3.5 35B-A3B Base was released on February 24, 2026, and Qwen3.5 0.8B on March 2, 2026.
Licensing and access
Open weights. Qwen3.5 35B-A3B Base is released as open weights under the Apache-2.0 license.
Yes. It is licensed under Apache 2.0, a permissive license that allows commercial use.
Basics
Qwen made Qwen3.5 35B-A3B Base, and it was released on February 24, 2026.
Yes. Qwen3.5 35B-A3B Base is released as open weights.
Generated from Vertical Tabs data, updated November 10, 2026.
More in Qwen
Similar models
7.4BOlmo Hybrid 7B Instruct
7.4BOlmo Hybrid 7B Think
MiniMax-M2.5-highspeed
30BNemotron Cascade 2 30B A3B
959MPaddleOCR-VL 1.5
MiniMax-M2.7-highspeed
Trendyol Asure 12BTrendyol Asure 12B
105BSarvam 105B
Step 3.5 Flash 2603Step 3.5 Flash 2603
30BSarvam 30B
4.3BNemotron 3 Content Safety
399B
Discussion
No comments yet
Be the first to share your thoughts.