DeepSeek V3-032: DeepSeek’s Newly Launched AI Model Beats GPT-4.5 and Gemini 2.0 Pro Experimental in Artificial Analysis Intelligence Index, Equals Grok 3
DeepSeek V3-032, DeepSeek’s newly launched AI model, outperforms GPT-4.5 and Gemini 2.0 Pro in the Artificial Analysis Intelligence Index, matching Grok 3 in performance.
The newly launched DeepSeek V3-032 model has gained significantly higher scores than the other AI models available in the market, beating GPT-4.5 (51) and Gemini 2.0 Pro Experimental (49). The Chinese AI company launched DeepSeek V3-032, an update to the DeepSeek V3 model, under an MIT license on March 24, 2025. It is an MoE (Mixture-of-Experts) language model designed with 671 billion parameters and 37 billion activated per token. DeepSeek V3-032 scored 53, same as Grok 3. However, it outperformed other models like Claude 3.7 Sonnet, DeepSeek V3, Qwen2.5 Max, GPT-4o and others in the Artificial Analysis Intelligence Index for non-reasoning models with seven valuations. OpenAI To Release Vision Feature ‘Advanced Vision Model’ on macOS, Will Allow AI To See and Understand Images.
DeepSeek V3-032 Delivers Scores Same as Grok 3
DeepSeek takes the lead: DeepSeek V3-0324 is now the highest scoring non-reasoning model
This is the first time an open weights model is the leading non-reasoning model, a milestone for open source.
DeepSeek V3-0324 has jumped forward 7 points in Artificial Analysis… pic.twitter.com/t7geJnQiBs
— Artificial Analysis (@ArtificialAnlys) March 25, 2025
(The above story first appeared on LatestLY on Mar 25, 2025 05:19 PM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).
