Only Activating 3B Parameters Is Comparable to GPT-4o. Qwen3 Updated Late at Night, and the First-hand Test Is Here!
In the early morning of July 29th, the Qwen team announced another major update - the already highly acclaimed Qwen3-30B-A3B model has ushered in a new version:Qwen3-30B-A3B-Instruct-2507.Machine learning enthusiast Vaibhav (VB) Srivastav was the first to share his feedback: "The latest Qwen3-30B-A3B-2507 runs extremely fast on a Mac equipped with MLX."
Focusing on the official data, this new model in non-thinking mode improves the ability to understand long texts to 256K.By activating only 3B parameters, you can achieve superb performance comparable to top closed-source models such as Gemini 2.5-Flash (non-thinking) and GPT-4o.at the same time,It has shown significant improvements in instruction following, logical reasoning, text comprehension, mathematics, science, programming, and tool usage.
Currently, "One-click deployment of Qwen3-30B-A3B-Instruct-2507" has been released to the OpenBayes public tutorial, and you can quickly experience the demo by cloning it with one click.We have conducted actual tests for everyone, asking whether there is a connection between the two extreme weather phenomena of heavy rain in many areas of Beijing and the typhoon landing in Shanghai. This model of non-thinking mode was tested and it quickly gave answers from multiple angles.
In addition, we have prepared surprise computing resource benefits for new users.Register with the invitation code "Qwen3-2507" to get 2 hours of dual-SIM A6000 usage time (resource valid for 1 month).The quantity is limited, don’t miss it!
Tutorial Link:
Demo Run
1. On the hyper.ai homepage, select the Tutorials page, choose One-click Deployment of Qwen3-30B-A3B-Instruct-2507, and click Run this tutorial online.
2. After the page jumps, click "Clone" in the upper right corner to clone the tutorial into your own container.
3. Select the NVIDIA RTX A6000-2 and PyTorch images. The OpenBayes platform has launched a new billing method: you can choose between "Pay as you go" or "Daily/Weekly/Monthly" according to your needs. Click "Continue." New users can register using the invitation link below to receive 4 hours of free RTX 4090 and 5 hours of free CPU time!
HyperAI exclusive invitation link (copy and open in browser):
https://go.openbayes.com/9S6Dr
4. Wait for resources to be allocated. The first clone will take about 2 minutes. When the status changes to "Running", click the jump arrow next to "API Address" to jump to the WebUI page. Please note that users must complete real-name authentication before using the API address access function.
Effect Demonstration
1. Extreme weather has been occurring frequently recently. After Beijing experienced a series of heavy rains, Shanghai was hit by a typhoon. Let's ask Qwen3-30B-A3B-Instruct-2507 whether there is any connection between the Shanghai typhoon and the Beijing heavy rains, and see what it says.
* After entering the API, if "Model" is not displayed in the upper left corner, it means the model is being initialized. Since the model is large, please wait about 2-3 minutes and refresh the page.
2. This version of the model is a new non-thinking mode model that objectively provides analysis from multiple perspectives.
The above is the tutorial recommended in this issue. Welcome everyone to experience it yourself~
Tutorial Link: