ModelsModels / Qwen
Qwen3-VL-30B-A3B-Instruct
Open weights · Apache-2.0by QwenReleased Sep 30, 2025
Qwen3-VL-30B-A3B-Instruct is made by Qwen. It takes text and images, with a context window of 262,144 tokens. On Tīrtha it takes text.
Start using Qwen3-VL-30B-A3B-Instruct in 3 steps
Read the docs- Get an API keyFree to create. It works for every model in the catalogue.Get an API key
- Add creditsPay as you go. Credits work on every model.Add credits
- Point your client at TīrthaAny OpenAI-compatible client works. Use your key and this base URL, and name the model.base_url: "https://api.tirtha.ai/v1" api_key: "YOUR_TIRTHA_KEY" model: "catalog/qwen3-vl-30b-a3b-instruct"
Price
Inputper 1M tokens
Outputper 1M tokens
Cached inputper 1M tokens
The live price, read from billing as this page loads. You pay only for the tokens you use, from your credits.
At a glance
MakerQwen
Acceptstext + image
Context262,144 tokens
Max output16,384 tokens
LicenceApache-2.0
ReleasedSep 30, 2025
Tested on Tīrtha Bench
Not on the public set yet: we have not run Qwen3-VL-30B-A3B-Instruct on the public 800-task tool-call set. The bench shows every model we have run so far.
What it supports
- Text inYes
- Image inYesText only on Tīrtha for now.
- StreamingYes
- Sampling settingsFixed todayThe maker suggests temperature 0.7, top p 0.8, top k 20, repetition penalty 1. Tīrtha fixes temperature at 0 today.
From the maker
- Model card: Qwen/Qwen3-VL-30B-A3B-Instruct on Hugging Face
- Last updated by the maker: Nov 26, 2025
Facts here are read from the maker's published card; claims and benchmarks on that card are the maker's own, not measured by Tīrtha.
Licence
Weights: Apache-2.0. Read the licence.