
How far up one can go, depends on how willing one is willing to climb.
Things were rather quiet, until things got disrupted. After OpenAI introduced ChatGPT, a war rages between tech companies. Those that saw the opportunity to commercialize Large Language Models, began to develop and launch their own AIs, upgrading with newer versions at an astonishing rate.
Life moves pretty fast in chatbot land.
But at first, things were focused on the West, with tech companies like OpenAI, Google, Meta, Apple and smaller ones like Perplexity, Anthropic and others. But since the East started to notice this trend, they started jumping into the bandwagon, betting their chances to thrive.
DeepSeek, for example, the company from China has what it calls the DeepSeek-V3, a multi-purpose AI model that rivals OpenAI GPT-4o and Meta LLaMA 3.1
Now, Alibaba is catching up fast, and managed to beat DeepSeek-V3 by a long margin.
The Chinese tech titan does this with what it calls the 'Qwen 2.5-Max' AI model.

"Qwen 2.5-Max outperforms ... almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B," Alibaba's cloud unit said in an announcement posted on its official WeChat account, referring to OpenAI and Meta's most advanced open-source AI models.
Alibaba said its Qwen 2.5-Max is par if not outperform its rivals in various benchmarks, including Arena-Hard, LiveBench, LiveCodeBench, and GPQA-Diamond. It also showed strong performance in other assessments such as MMLU-Pro.
"We have been building Qwen2.5-Max, a large MoE LLM pretrained on massive data and post-trained with curated SFT and RLHF recipes," Alibaba Cloud announced in a blog post, emphasizing that the model has been trained with more than 20 trillion tokens.

And as for its abilities, include the ability to perform a number of text and image analysis tasks, parse files, understand videos, and count objects in images.
The AI can also control a PC, an ability similar to OpenAI’s recently launched Operator.
"Now Qwen2.5-Max is available in Qwen Chat, and you can directly chat with the model, or play with artifacts, search, etc.,” Alibaba said.
Here, Qwen 2.5-Max is able to demonstrate competitive results against industry leaders like GPT-4o and Claude-3.5-Sonnet in tests of advanced reasoning and knowledge, despite not being a purposefully built reasoning model.
What this means, Alibaba's Qwen2.5-Max, which is a multi-purpose AI, is more like DeepSeek-V3 or GPT-4o than proper reasoning models like DeepSeek-R1, or OpenAI o1, or Google Gemini 2.0 Flash Thinking, for example.
But here's the thing, Qwen2.5-Max actually outperformed DeepSeek-R1, whose success sent Nvidia’s stock plunging.
Just like DeepSeek and many other Chinese tech companies that are deprived from Western-made computer chips, Alibaba created Qwen2.5-Max using a mixture-of-experts architecture that requires significantly fewer computational resources than traditional approaches.
The timing of these back-to-back Chinese AI releases has deepened America's anxiety about U.S. technological supremacy.
With both announcements came during President Donald Trump’s first week back in office, prompting questions about the effectiveness of U.S. chip export controls meant to slow China’s AI advancement.
It's worth noting that Alibaba and DeepSeek are just two of more than handful of tech companies from China entering this arms race. Others include Baidu, a tech heavyweight competitor regarded as the creator of China's first equivalent to ChatGPT, and the country's most valuable internet company Tencent.
Qwen2.5-VL, another LLM that is part of the Qwen that can be tested through Alibaba’s Qwen Chat app, available for download from AI developer platform Hugging Face, can analyze charts and graphics, extract data from scans of invoices and forms, and "comprehend" multiple-hours-long videos, the Qwen team said.
Qwen2.5-VL can also recognize "IPs from film and TV series, as well as a wide variety of products."
Read: DeepSeek, The Chinese AI That Makes Silicon Valley Nervous And The U.S. Concerned