Market
Which AI model will have the highest user satisfaction rating in 2025?
2
Alibaba Group•OpenAI•GPT•Meta•Llama•Alibaba Cloud•WeChat•SFT•RLHF•Alibaba•Arena Hard•LiveBench•LiveCodeBench•Qwen
Resolution / Starting Odds
Qwen2.5-Max • 25%
DeepSeek-V3 • 25%
GPT-4o • 25%
Llama-3.1-405B • 25%
Results from a major tech survey or user satisfaction study
Story
Alibaba Cloud's Qwen2.5-Max AI Model Claims to Surpass DeepSeek-V3, GPT-4o
Jan 29, 2025, 11:14 AM
Alibaba Group's cloud computing unit has unveiled Qwen2.5-Max, a new artificial intelligence model that the company claims surpasses leading AI models including DeepSeek-V3, OpenAI's GPT-4o, and Meta's Llama-3.1-405B. Announced on Alibaba Cloud's official WeChat account, Qwen2.5-Max is a large Mixture-of-Experts (MoE) language model pretrained on massive data and post-trained with curated Supervised Fine-Tuning (SFT) and Reinforcement Learning with Human Feedback (RLHF) techniques. According to Alibaba, the model outperforms competitors in benchmarks such as Arena Hard, LiveBench, LiveCodeBench, and GPQA-Diamond. The company expects future versions of Qwen to achieve even greater performance with improved post-training methods. The release signifies Alibaba's response to the intense competition in the AI sector, particularly following the recent success of DeepSeek-V3.
View original story
Similar markets
DALL-E 3 • 25%
Stable Diffusion • 25%
Other • 25%
Janus-Pro-7B • 25%
OpenAI o1 • 25%
Other • 25%
DeepSeek-R1 • 25%
Google's AI model • 25%
Other • 25%
DeepSeek-R1 • 25%
OpenAI's o1 • 25%
Google's Gemini • 25%
Other • 25%
Llama 4 • 25%
No clear leader • 25%
Another AI model • 25%
Phi-4 • 25%
Gemini Pro • 25%
Llama 3.3 • 25%
Other • 25%
Another AI model • 25%
OpenAI's o1 • 25%
DeepSeek-R1 • 25%
No clear leader • 25%
Other AI models lead • 25%
DeepSeek V3 leads • 25%
OpenAI leads • 25%
Meta leads • 25%
Gemini Pro • 25%
Phi-4 • 25%
Llama 3.3 • 25%
Other • 25%
DeepSeek-V3 • 25%
Qwen2.5-Max • 25%
Llama-3.1-405B • 25%
GPT-4o • 25%
DeepSeek-V3 • 25%
Qwen2.5-Max • 25%
Llama-3.1-405B • 25%
GPT-4o • 25%
Other • 25%
DeepSeek R1 • 25%
Another existing model • 25%
A new model • 25%
Other • 25%
Llama 4 • 25%
GPT-5 • 25%
Anthropic's Claude • 25%