Google's new Gemini model now holds the No. 1 spot in the Chatbot Arena across all domains

The competition between OpenAI and Google is heating up day by day. The large language models from both these companies are taking turns to top the AI benchmarks.

In the popular Chatbot Arena leaderboard, ChatGPT-4o (20241120) became the No. 1 model on Nov. 21st, surpassing the Gemini-Exp-1114 model, which was released on Nov. 15th. Google today announced the release of the Gemini-Exp-1206 experimental model, which beats ChatGPT-4o (20241120) in the Chatbot Arena.

Today’s the one year anniversary of our first Gemini model releases! And it’s never looked better.

Check out our newest release, Gemini-exp-1206, in Google AI Studio and the Gemini API!https://t.co/CCQwBxm8Yu

— Jeff Dean (@🏡) (@JeffDean) December 6, 2024

Gemini-Exp-1206 is the No. 1 model overall and tied with OpenAI"s o1 in the coding category. Gemini-Exp-1206 is now No. 1 in the following categories:

  • Overall with Style Control
  • Hard Prompts
  • Hard Prompts with Style Control
  • Coding
  • Math
  • Creative Writing
  • Instruction Following
  • Longer Query
  • Multi-Turn

The Gemini-exp-1206 model is now available both in Google AI Studio and the Gemini API for developers.

On a related note, Meta today released the Llama 3.3 70B open-source model that delivers class-leading performance for text-based use cases. Meta claims that Llama 3.3 70B is available at a fraction of the inference cost when compared to other competing closed-source models.

Meta claims that Llama 3.3 improvements were mainly because of a new alignment process and progress in online RL techniques. It is important to note that this new 70B model delivers similar performance to Llama 3.1 405B, but it can even run locally on common developer workstations.

As we continue to explore new post-training techniques, today we"re releasing Llama 3.3 — a new open source model that delivers leading performance and quality across text-based use cases such as synthetic data generation at a fraction of the inference cost. pic.twitter.com/BNoV2czGKL

— AI at Meta (@AIatMeta) December 6, 2024

This new Llama 3.3 70B model is available from Hugging Face. Soon it will be available for deployment through Azure and other leading cloud platforms. With the availability of powerful and affordable open-source models like Llama 3.3, the future of AI development looks promising and accessible to a wider developer community.

Report a problem with article
Next Article

Windows 11 build 26120.2510 brings Recall and other AI features to AMD and Intel Copilot PCs

Previous Article

Save 80% on a lifetime subscription to Headway Premium