It’s my great honor to have just received the test-of-time award at #ACL2025 from @aclmeeting for our paper “Effective Approaches to Attention-based Neural Machine Translation” (joint work with @hyhieu226 and @chrmanning at EMNLP 2015). It’s the immediate second “attention”
@lmthang
-
DeepMind Team Finalizes IMO 2025 Preparation in London
By
–
Right before #imo2025, together with colleagues from Mountain View, NYC, Singapore, etc, we all gathered at @GoogleDeepMind headquarter in London for our final push for IMO. I believe that week was when all magic happened! We put all individual recipes (that we figured out
-
Google Launches Advanced Gemini with Deep Think for Mathematicians
By
–
And I am excited that we will very soon be bringing this advanced version of Gemini with Deep Think first to the hand of mathematicians and subsequently more people. See more details in the blog:
-
Gemini advances multi-step reasoning with novel reinforcement learning techniques
By
–
This year was a major paradigm shift, where we can solve problems end to end in natural language. With novel reinforcement learning techniques, we are able to train an advanced Gemini model on multi-step reasoning proof data, which advances the model's capabilities in terms of
-
Gemini Deep Think Wins Gold at International Mathematical Olympiad
By
–
Very excited to share that an advanced version of Gemini Deep Think is the first to have achieved gold-medal level in the International Mathematical Olympiad! , solving five out of six problems perfectly, as verified by the IMO organizers! It’s been a wild run to lead this
-
Catch a Mole: Challenging Codeforces Problem Demo
By
–
Also check out this cool and fun demo on "catch a mole", a challenging problem from Codeforces! https://t.co/x9ksPEWjCx
— Thang Luong (@lmthang) 23 mai 2025Also check out this cool and fun demo on "catch a mole", a challenging problem from Codeforces!
-
Gemini 2.5 Pro achieves state-of-the-art across math coding multimodality
By
–
Not only math, Gemini 2.5 Pro with #DeepThink also achieved state-of-the-art performances for coding (on LiveCodeBench v6), multimodality (on MMMU), and much more! Check out all other #GoogleIO amazing developments with Gemini 2.5 from @GoogleDeepMind and @Google at
-
Gemini 2.5 Pro Deep Think Mathematical Proof Analysis
By
–
Attached is a full proof by Gemini 2.5 Pro #DeepThink with our experts' comments https://
drive.google.com/file/d/1PaKXo4
0u_c4Ti7k5-3gz8b1cjp92U7ED/view?usp=sharing
…. Here I quote a few important moments in the proof:
(a) Expert: The main part of the solution begins with a "proof by contradiction", which is a reasonable choice considering -
Gemini 2.5 Pro Achieves 49.4% on USAMO 2025 Math Benchmark
By
–
As highlighted yesterday at #GoogleIO, Gemini 2.5 Pro with #DeepThink mode achieved a new state-of-the-art performance of 49.4% on USAMO 2025, an extremely challenging Math benchmark for AI that requires detailed proof writing.
— Thang Luong (@lmthang) 22 mai 2025
I was analyzing the model outputs with our math… https://t.co/hd8XjbYVA7 pic.twitter.com/sU5uzhf2IzAs highlighted yesterday at #GoogleIO, Gemini 2.5 Pro with #DeepThink mode achieved a new state-of-the-art performance of 49.4% on USAMO 2025, an extremely challenging Math benchmark for AI that requires detailed proof writing. I was analyzing the model outputs with our math
-
Google Launches AI Ultra Plan with VIP Access to Advanced Tools
By
–
We didn't propose the Google AI Ultra plan $249.99/month, but hopefully it will be well worth it with your VIP access to Google AI (DeepThink, Veo3, Project Mariner, etc) 🙂