Chinese Giant Alibaba Claims to Have Developed Its Own AI—And It’s Beating ChatGPT and DeepSeek!

Published on
Read : 3 min
Qwen2.5 Max From Alibaba
Chinese Giant Alibaba Claims to Have Developed Its Own AI—And It’s Beating ChatGPT and DeepSeek! | The Daily Galaxy --Great Discoveries Channel

The AI war has just exploded. In a week already dominated by China’s DeepSeek unveiling its game-changing R1 model, Alibaba has now thrown down the gauntlet, claiming its new Qwen2.5-Max is one of the most powerful and advanced AI models ever created.

Not only does Alibaba assert that Qwen2.5-Max can go toe-to-toe with industry leaders like Claude-3.5-Sonnet from Anthropic, it claims the model surpasses GPT-4o from OpenAI and Llama-3.1-405B from Meta in multiple key benchmarks.

But what exactly does this mean? Is this the moment China takes the lead in the AI race? And how did DeepSeek’s shock launch spark this sudden AI arms race?

Qwen2.5-Max: The AI That Could Dethrone OpenAI?

Alibaba’s Qwen2.5-Max, presented on January 29, is a partially open-source AI model that has been engineered to compete at the highest level of artificial intelligence. Unlike some of its rivals, Alibaba is taking a hybrid approach, blending open-source transparency with proprietary advancements.

According to data released by Alibaba, Qwen2.5-Max has been rigorously tested across five major AI benchmarks:

  • Arena-Hard – Evaluates AI’s ability to align with human preferences, a crucial factor for AI assistant-like usability.
  • LiveBench – A general intelligence test that measures an AI model’s adaptability to a wide range of queries.
  • LiveCodeBench – Assesses AI’s coding capabilities, a key performance area for enterprise AI applications.
  • GPQA-Diamond – Measures AI’s knowledge and reasoning in science fields such as biology, physics, and chemistry.
  • MMLU-Pro – An advanced AI benchmark that tests AI’s college-level understanding across multiple disciplines.

Alibaba claims that Qwen2.5-Max is on par with Claude-3.5-Sonnet, and dominates GPT-4o, DeepSeek-V3, and Llama-3.1-405B in most of these categories. However, it falls slightly behind Claude-3.5-Sonnet in LiveCodeBench and trails in GPQA-Diamond, a benchmark that tests deep knowledge of scientific disciplines.

This means that while Qwen2.5-Max isn’t an undisputed champion, it is one of the strongest challengers OpenAI, Anthropic, and Meta have ever faced.

The Deepseek Shockwave: How China’s AI Industry Set Wall Street on Fire

The unveiling of DeepSeek-V3 earlier this week was already a massive disruptor. DeepSeek’s chatbot quickly became the #1 most downloaded free app on Apple’s U.S. App Store, beating ChatGPT in popularity. This unexpected explosion in interest sent a wave of panic through global financial markets.

The result? A historic $1 trillion loss across major U.S. tech stocks in just a few days.

Biggest victims of the DeepSeek Effect:

  • Nvidia suffered the worst single-day market loss in U.S. history, dropping 60% in value.
  • Microsoft, Meta, and Google saw hundreds of billions wiped off their stock prices as investors reeled from the sudden AI power shift.
  • OpenAI and Anthropic, the U.S.’s top AI contenders, are now racing to respond to China’s unexpected advancements.

Why DeepSeek matters

DeepSeek’s R1 model made headlines not just for its competitive performance, but for its affordability and efficiency. Some experts believe it outperforms OpenAI’s GPT-4 in key areas while being significantly cheaper to operate.

The tech world did not see this coming, and the result was panic on Wall Street as investors rushed to reassess the AI landscape.

China’s AI Blitz: Alibaba, Bytedance, and Deepseek Push Back

The AI showdown is far from over. Following DeepSeek’s record-breaking debut, China’s other AI titans immediately sprang into action:

  • ByteDance (TikTok’s parent company) rushed to update its AI model, Doubao, in response to DeepSeek’s sudden rise.
  • Alibaba, likely feeling the pressure, unveiled Qwen2.5-Max ahead of schedule to keep up with the escalating competition.

The takeaway? The AI war is accelerating faster than anyone anticipated.

The AI Battleground: Who Leads Now?

While Qwen2.5-Max is an impressive contender, it still faces fierce competition from multiple fronts. Here’s a breakdown of where things stand:

AI ModelTop StrengthsWeaknesses
Qwen2.5-Max (Alibaba)Excels in Arena-Hard and LiveBench, strong in general intelligence and alignmentSlightly weaker in LiveCodeBench, trails in GPQA-Diamond (science knowledge)
DeepSeek-V3Extremely cost-effective, performs well in general AI tasksYet to prove superiority in specialized tasks
Claude-3.5-Sonnet (Anthropic)Strongest in LiveCodeBench, excellent at human-aligned responsesLess aggressive in AI arms race, limited access
GPT-4o (OpenAI)Massive global user base, proven leader in multi-modal AIBeing challenged by newer, faster, and cheaper models
Llama-3.1-405B (Meta)Strong performance across benchmarksLacks major breakthroughs against competition

The real question now is: how will OpenAI, Microsoft, and Google respond?

Follow us on Google NewsDaily Galaxy - Support us by adding us to your Google News favorites.

Leave a Comment