Research Scientist @GoogleDeepMind | Agents and Code in Gemini
London, England
Joined March 2012
- Tweets527
- Following7K
- Followers1.7K
- Likes3.1K
Abhinav Gupta retweeted
Happening now!! Drop by if you are interested in test-time learning or making models perform better with limited data!!
📍Pavilion 3 , P3 - #206
@iclr_conf #ICLR2026 @EdinburghNLP
Excited to share that I will be presenting my 1st year PhD work at #ICLR2026 @iclr_conf next week in Rio 🇧🇷
❓In-Context Learning (ICL) has been a widely used paradigm to adapt LLMs for a task with a few examples or shots. But does it also work for multimodal examples? 🧵
Abhinav Gupta retweeted
Been waiting a long time to share this one. Meet Gemini 3 Pro. It’s our most intelligent multimodal model that’s deeply capable.
xitter.cf/sundarpichai/status/19…
Introducing Gemini 3 ✨
It’s the best model in the world for multimodal understanding, and our most powerful agentic + vibe coding model yet. Gemini 3 can bring any idea to life, quickly grasping context and intent so you can get what you need with less prompting.
Find Gemini 3 Pro rolling out today in the @Geminiapp and AI Mode in Search. For developers, build with it now in @GoogleAIStudio and Vertex AI.
Excited for you to try it!
Abhinav Gupta retweeted
The secret behind Gemini 3?
Simple: Improving pre-training & post-training 🤯
Pre-training: Contra the popular belief that scaling is over—which we discussed in our NeurIPS '25 talk with @ilyasut and @quocleix—the team delivered a drastic jump. The delta between 2.5 and 3.0 is as big as we've ever seen. No walls in sight!
Post-training: Still a total greenfield. There's lots of room for algorithmic progress and improvement, and 3.0 hasn't been an exception, thanks to our stellar team.
Congratulations to the whole team 💙💙💙
Abhinav Gupta retweeted
Excited to launch Google Antigravity, our next generation agentic IDE, now powered by Gemini 3!
Abhinav Gupta retweeted
We are launching Gemini-pro-exp-03-25 which is our most capable model and generally useful on a wide vareity of real-world tasks. It's #1 on LMArena and SOTA on a wide set of benchmarks. This is a massive Gemini wide effort and I am incredibly proud of the team behind this. 🚀🚀
BREAKING: Gemini 2.5 Pro is now #1 on the Arena leaderboard - the largest score jump ever (+40 pts vs Grok-3/GPT-4.5)! 🏆
Tested under codename "nebula"🌌, Gemini 2.5 Pro ranked #1🥇 across ALL categories and UNIQUELY #1 in Math, Creative Writing, Instruction Following, Longer Query, and Multi-Turn!
Massive congrats to @GoogleDeepMind for this incredible Arena milestone! 🙌
More highlights in thread👇
Abhinav Gupta retweeted
The accusations/obsessions over DeepSeek using H100 sound like a rich kids team got outplayed by a poor kids team, who weren't even allowed shoes, and now the rich kids are demanding an investigation into whether shoes were used instead of training harder to improve themselves.
Abhinav Gupta retweeted
Deep Dive is now my favorite podcast. The more I listen the more I feel like I'm becoming friends with the hosts and I think this is the first time I've actually viscerally liked an AI. Two AIs! They are fun, engaging, thoughtful, open-minded, curious. ok i'll stop now.
Abhinav Gupta retweeted
So true. My offer from UofT was considerably lower than any US school. This has made recruiting good students from the US very difficult for them.
Abhinav Gupta retweeted
Still 3 more days to submit! We also accept recent work submitted/accepted to conferences (e.g., ICLR, EMNLP, AAMAS, NeurIPS) and other workshops! 🤖📡 #NeurIPS2020
Submissions are open for Emergent Communication @ NeurIPS 2020 🥳 sites.google.com/view/emecom…
@filangelos @aggielaz @j_foerst @backpropper @mnoukhov @BullardKalesha
Abhinav Gupta retweeted
Conald Trump is a truly awful and evil man. His behavior was disgusting the entire 90 minutes and he wouldn’t follow the rules he agreed to.
The rules were clear. Trump broke them constantly as he does with most things.
A disgrace. #Debates2020
Abhinav Gupta retweeted
related to previous tweet, my (controversial?) take: we (CS) put too much emphasis on creating novel algorithms. this is dumb. first of all, most of them aren't really novel. second, we have so many good existing ones already. it's stupid to be allowed to use an algo only once.
Abhinav Gupta retweeted
Read this article today on tips for reading scientific papers. I find myself wanting to read every word and understand every figure, but I have to learn to read more strategically! @AcademicChatter #phdchat #phdlife #gradschool sciencemag.org/careers/2016/…
Abhinav Gupta retweeted
an infinite loop. nicely done Santa Clara County PUBLIC HEALTH reddit.com/r/ProgrammerHumor…
Abhinav Gupta retweeted
Probably already mentioned but a suggestion for @NeurIPSConf is to keep reviewers anonymous during post rebuttal discussion. If not, known "experts" or more senior researchers could dominate the conversations and junior researchers wouldn't want to disagree to hurt relationships.
Abhinav Gupta retweeted
Submissions are open for Emergent Communication @ NeurIPS 2020 🥳 sites.google.com/view/emecom…
@filangelos @aggielaz @j_foerst @backpropper @mnoukhov @BullardKalesha
Abhinav Gupta retweeted
Rebuttals are borderline pointless without being able to see a revised manuscript
Abhinav Gupta retweeted
Best advice I got in grad school: technical difficulty isn't predictive of research merit.
Some important problems require little technical novelty and some work's only contribution is the difficulty of its creation.
Hopefully @NeurIPSConf reviewers will learn this one day.
Abhinav Gupta retweeted
This brilliantly written summary by @ah__cl should be mandatory reading for anyone working in AI who intends to use the C-word #compositionality
thegradient.pub/how-to-stop-…
Abhinav Gupta retweeted
hello world! this is a twitter handle of the machine learning lab, CILVR, at @nyuniversity. see our homepage at cilvr.nyu.edu and in particular our newly created blog at wp.nyu.edu/cilvr/cilvr-blog/