OpenAI has always had a good counter to anything we ship, magically always on the same day. But I'm not very impressed with Santa mode as a counter to Gemini 2.0, doesn't quite have the same gravitas.
Jonas Adler
260 posts
Jonas Adler
@JonasAAdler
Research Scientist, DeepMind.
AlphaFold, Gemini.
- Competing with ourselves is getting a bit boring🚨Breaking: @GoogleDeepMind’s latest Gemini-2.5-Pro is now ranked #1 across all LMArena leaderboards 🏆 Highlights: - #1 in all text arenas (Coding, Style Control, Creative Writing, etc) - #1 on the Vision leaderboard with a ~70 pts lead! - #1 on WebDev Arena, surpassing Claude
- Replying to @tszzlBut I thought the tradition was that you ship the same day as us :(
- Pace. Focus. Intensity.Our latest Gemini 2.5 Pro update is now in preview. It’s better at coding, reasoning, science + math, shows improved performance across key benchmarks (AIDER Polyglot, GPQA, HLE to name a few), and leads @lmarena_ai with a 24pt Elo score jump since the previous version. We also
- Few people use Google search because we hyped it or did fancy commercials. My goal for Gemini is the same.The under-hype over-deliver phenomenon in action 👇🏻
- Deep learning is clearly hitting a wallAnnouncing AlphaFold 3: our state-of-the-art AI model for predicting the structure and interactions of all life’s molecules. 🧬 Here’s how we built it with @IsomorphicLabs and what it means for biology. 🧵 dpmd.ai/3URDiNo
00:00 - Such a good model :) We really had quite a lot of catch-up to do, but now we're in flip-flop mode and seemingly ahead more often than behind. Amazing!Big news on Chatbot Arena 🔥 The new @GoogleDeepMind model gemini-exp-1206 is crushing it, and the race is heating up. Google is back in the #1 spot 🏆overall and tied with O1 for the top coding model! Highlights (improvement since gemini-exp-1121 in parentheses) - First
- We had a lot of catch up to do, but we've done it. Amazing work from the team!Exciting News from Chatbot Arena! @GoogleDeepMind's new Gemini 1.5 Pro (Experimental 0801) has been tested in Arena for the past week, gathering over 12K community votes. For the first time, Google Gemini has claimed the #1 spot, surpassing GPT-4o/Claude-3.5 with an impressive
- Replying to @NandoDFIndeed two years ago when we were wrapping up AlphaFold3 I moved over to Gemini because I thought it was the best way to top AlphaFold and make things better. At that time Gemini was quite decidedly "off-frontier" and we've had to work really hard to become relevant and useful.
- Our papers are finally out in hard copy, and with a fantastic cover illustration at that. I'll be buying an issues at least!The #AlphaFold 2 papers on the methods and human proteome predictions are out today in hard copy in @Nature! A really proud moment to see our work featured with a fantastic image on the front cover of the issue: dpmd.ai/proteinpower
- My favourite part of this might be that deep down into the report we have some crazy math numbers for a math focused model. It wasn't long ago that Minerva scored 50%. The best model in 2021 scored about 7%. 90% is on part with IMO gold medalists given 1h per problem.
Gemini 1.5 Model Family: Technical Report updates now published In the report we present the latest models of the Gemini family – Gemini 1.5 Pro and Gemini 1.5 Flash, two highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information - Give DeepMind a good leaderboard and we'll climb it. So many tiny improvements went into making this happen. When you're in the midst of it you get lost in the daily details, but then you look look up and notice the amazing progress that happens on a month by month basis.Hundreds of researchers make thousands of discoveries. Most researchers focus only on a specific part of the problem, and yet when all those discoveries are put together, it compounds. Pretty incredible to witness.
- Replying to @suchenzangEntropy always increases, ergo life cannot exist. Checkmate living things.
- Shipped! It did take some time, probably in no small part since the team tried to clean up all the messy code I wrote.The #AlphaFold 3 model code and weights are now available for academic use. We @GoogleDeepMind are excited to see how the research community continues to use AlphaFold to address open questions in biology and new lines of research. github.com/google-deepmin…
