Pinned Tweet
My PhD thesis "Deep Learning with Graph-Structured Representations" is now available for download: hdl.handle.net/11245.1/1b63b9… -- It covers a range of emerging topics in Deep Learning: from graph neural nets (and graph convolutions) to structure discovery (objects, relations, events)
43
599
3,167
We just released an update to Omni 🎬
This is a much bigger step than 1.0->1.1 — give it a try!
This is a much bigger step than 1.0->1.1 — give it a try!
Meet Gemini Omni 1.2 Flash ⚡️
We’ve made the model more cost-effective, bumped the video generation and editing limit up to 15 seconds, and made updates to how scenes hold together over time. We've also given character speech an upgrade, reduced visual glitches like jump cuts, stays even truer to any reference assets you upload.
So… what does all of that actually mean for you?
It gives you more room to tell your story. With up to 15 seconds to work with, you have more flexibility to create what you need — videos look more cinematic and smooth, offering stronger consistency that keeps objects anchored from start to finish. And when characters speak, the dialogue sounds more fluid and natural.
Creators and everyday users can start directing scenes and multi-turn edits in @FlowbyGoogle, generating 15-second clips at no additional cost in Google Vids, or creating 15-second videos in the @GeminiApp. Plus, for developers and businesses building with the API, we've dropped the price by 30% from the previous generation, making it available at just $0.07 per second for 720p video.
Watch the model in action as it turns a single style frame into a 4-shot, 15-second macro science explainer of a living beehive ↓
3
5
52
2,758
Thomas Kipf retweeted
The Ocean Cleanup is now $40M/year short of what’s needed to stop up to 1/3rd of all plastic flowing into the oceans by 2030.
That’s less than 1% of what America spends on Halloween decorations each year.
That’s less than 1% of what America spends on Halloween decorations each year.
214
777
5,174
148,307
Thomas Kipf retweeted
We’ve released AA-Omniscience v1.1, updating our knowledge and reasoning benchmark with a refreshed public set, removal of saturated questions, and ambiguity fixes
AA-Omniscience is our proprietary benchmark measuring model knowledge and hallucinations, including whether models now an answer. Our primary metric, Net Score, rewards correct answers, penalizes incorrect answers, and treats abstentions neutrally.
AA-Omniscience was updated to v1.1 as part of the Artificial Analysis Intelligence Index v5 release.
Key changes in AA-Omniscience v1.1:
➤ Updated 3,000-question dataset: Down from 6,000 questions, with previously public and saturated questions removed, errors and ambiguities corrected, and new questions added
➤ Newly calibrated 300-question public set: AA-Omniscience v1.1 has a new public set on @huggingface calibrated to match the difficulty of the private set. At release, public and private scores have a 0.99 Pearson correlation and 0.99 Spearman rank correlation for AA-Omniscience Net Score, with a median difference of 0.0 points. Accuracy and hallucination rate also have 0.99 correlations, with median differences of -0.07 and -0.29 percentage points, respectively
➤ Slightly updated domain distribution: Questions span the same 6 domains and 42 topics as version 1, with a revised distribution: Science, Engineering and Mathematics (600), Software Engineering (600), Finance (450), Health (450), Humanities and Social Sciences (450), and Law (450)
Key results for AA-Omniscience v1.1:
➤ Rank order largely preserved: AA-Omniscience v1.1 has a 0.98 Spearman rank correlation with v1. The recently released Gemini 4 Carbon (high) leads with a Net Score of 54, followed by Claude Opus 5.5 (48, max with fallback) and Claude Fable 5.1 (46, max with fallback). Muse Spark Open (xhigh) is the leading open weights model, 28 points behind Gemini 4 Carbon
➤ Scores average ~2 points lower on the refreshed set: Models score 1.9 points lower on Net Score on average compared to v1, with a median decrease of 0.6 points. Some models see much larger declines, particularly MiniMax M3.1 Flash, down 22 points (16 to -6), and MiMo V2.6 Pro, down 13 points (8 to -5)
➤ Slightly lower accuracy: AA-Omniscience Accuracy measures the percentage of questions the model answers correctly. Models answer a median of 1.5 percentage points fewer questions correctly on average compared to v1. Claude Fable 5.1 (max with fallback) has the highest accuracy at 68%, followed by Claude Opus 5.5 (max with fallback) at 66%
➤ Hallucination rates decrease: AA-Hallucination measures how often models choose to answer when they do not know the answer. The median Hallucination Rate falls 1.4 percentage points, with Gemini 4 Argon (high) recording the lowest hallucination rate at 4%
AA-Omniscience is our proprietary benchmark measuring model knowledge and hallucinations, including whether models now an answer. Our primary metric, Net Score, rewards correct answers, penalizes incorrect answers, and treats abstentions neutrally.
AA-Omniscience was updated to v1.1 as part of the Artificial Analysis Intelligence Index v5 release.
Key changes in AA-Omniscience v1.1:
➤ Updated 3,000-question dataset: Down from 6,000 questions, with previously public and saturated questions removed, errors and ambiguities corrected, and new questions added
➤ Newly calibrated 300-question public set: AA-Omniscience v1.1 has a new public set on @huggingface calibrated to match the difficulty of the private set. At release, public and private scores have a 0.99 Pearson correlation and 0.99 Spearman rank correlation for AA-Omniscience Net Score, with a median difference of 0.0 points. Accuracy and hallucination rate also have 0.99 correlations, with median differences of -0.07 and -0.29 percentage points, respectively
➤ Slightly updated domain distribution: Questions span the same 6 domains and 42 topics as version 1, with a revised distribution: Science, Engineering and Mathematics (600), Software Engineering (600), Finance (450), Health (450), Humanities and Social Sciences (450), and Law (450)
Key results for AA-Omniscience v1.1:
➤ Rank order largely preserved: AA-Omniscience v1.1 has a 0.98 Spearman rank correlation with v1. The recently released Gemini 4 Carbon (high) leads with a Net Score of 54, followed by Claude Opus 5.5 (48, max with fallback) and Claude Fable 5.1 (46, max with fallback). Muse Spark Open (xhigh) is the leading open weights model, 28 points behind Gemini 4 Carbon
➤ Scores average ~2 points lower on the refreshed set: Models score 1.9 points lower on Net Score on average compared to v1, with a median decrease of 0.6 points. Some models see much larger declines, particularly MiniMax M3.1 Flash, down 22 points (16 to -6), and MiMo V2.6 Pro, down 13 points (8 to -5)
➤ Slightly lower accuracy: AA-Omniscience Accuracy measures the percentage of questions the model answers correctly. Models answer a median of 1.5 percentage points fewer questions correctly on average compared to v1. Claude Fable 5.1 (max with fallback) has the highest accuracy at 68%, followed by Claude Opus 5.5 (max with fallback) at 66%
➤ Hallucination rates decrease: AA-Hallucination measures how often models choose to answer when they do not know the answer. The median Hallucination Rate falls 1.4 percentage points, with Gemini 4 Argon (high) recording the lowest hallucination rate at 4%
94
50
898
203,367
Gemini 4 Carbon is here!
It’s been a significant improvement for my work overall
It’s been a significant improvement for my work overall
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys 9. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.
We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards.
As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
0
1
24
1,859
Legendary. First found and fixed by Gemini 3.7 Flash
Hi all, I have a new result I'm excited to share. It turns out that *all existing implementations* of softmax and cross-entropy (that's JAX, PyTorch, textbooks, etc) have a subtle bug --- two bugs, actually. A simple rewrite fixes both, and eliminates "slingshots" when grokking:
5
7
139
20,886
Thomas Kipf retweeted
We added Gemini 4 Argon to our AutoResearchExam Pareto curve. It is 2-3 times cheaper compared to the other frontier models for almost the same ML engineering performance. Congratulations to the GDM team.
14
28
347
45,932
Thomas Kipf retweeted
As a German, it is a tragedy: in a parallel universe this would be Germany’s moment, as it requires so much that Germany has traditionally excelled at: industrial capacity, engineering, world-leading physics, design and manufacturing excellence. That was exactly what ‘made in Germany,’ stood for.
Jensen Huang has a message for Germany
104
221
2,026
120,582
Thomas Kipf retweeted
We’re rolling out Gemini 4 Argon in Antigravity to paid Google AI Pro subscribers today. Demand has been incredible and we've scaled capacity deliberately to get here. We’ve rolled out a separate weekly-only Argon quota for Pro on top of the existing Gemini bucket for 3.8 Flash so you have the best of Gemini for all your tasks.
1,071
1,328
15,164
1,507,481
Argon 🫶 Omni
Argon can also plan effectively in world models!
Here it went through an iterated loop of 1) writing out plans for how to achieve a goal 2) using an omni world model to simulate consequences of actions 3) reflect on produced rollouts to modify the plan. This is the resulting ls in our blog - developers.googleblog.com/gemini-4-argon…
1
0
18
2,305
Frontier models now enable zero-shot robotic control and other robotics use cases. Super excited about Gemini 4 Argon’s capabilities in this space.
Argon is very capable model for agentic robotics use cases. We've put together number of examples across control, sim from real, value prediction, hindsight labeling, etc: developers.googleblog.com/gemini-4-argon… Excited for folks to try it.
1
0
24
2,182
Thomas Kipf retweeted
Excited to launch Gemini 4 Argon, designed to help with hardest coding and knowledge work tasks with its frontier capabilities..
For developers and enterprises, it’s available in the paid tier of the Gemini API and the Gemini platform, for Ultra subscribers in Google Antigravity, and Ultra and Pro subscribers in the Gemini app and Search.
Read more: blog.google/innovation-and…
For developers and enterprises, it’s available in the paid tier of the Gemini API and the Gemini platform, for Ultra subscribers in Google Antigravity, and Ultra and Pro subscribers in the Gemini app and Search.
Read more: blog.google/innovation-and…
64
68
881
57,649
spot on
European households are sitting on €10–12 trillion in cash and bank savings. Invested, that money could power European growth. A stronger European economy means a better trading partner for America and a more capable NATO ally.
So why isn't that money invested? Savers don't believe companies in Europe can earn enough to justify the risk.
Mario Draghi's competitiveness report named the cause: Europe has piled rule upon rule while its productivity engine has stalled. Small businesses agree. They call regulation their biggest obstacle.
Making investing simpler, as Brussels plans, is a start. But savers follow growth. So, how do we get there?
Energy abundance, faster permits, rewards for investing in growing firms, and an end to overlapping reporting rules that turn mid-sized firms into reporting shops.
And one test for every new rule: Would a competent founder still choose to build and grow in Europe?
Make Europe a better bet, and the capital will follow.
Read my op-ed in @rapporteur_news ↓ rapporteur.com/opinion/europe…
1
0
13
6,014
Thomas Kipf retweeted
I work at Google DeepMind and yes I have tried the model known externally as “Argon”.
I can say this:
I don’t reach for Opus anymore for any work (like I did when 3.8 Flash was our top model).
I hope that it will soon be generally available for everyone to enjoy!
I can say this:
I don’t reach for Opus anymore for any work (like I did when 3.8 Flash was our top model).
I hope that it will soon be generally available for everyone to enjoy!
I have a simple question: Did anyone actually try Gemini 4 Argon?
314
128
6,241
708,052
That’s a solid recipe for developing RSI (repetitive strain injury)
Gemini 4 is out. I personally spent months carefully setting the parameters by hand. Around parameter 1 trillion I started getting a little sloppy, so if you notice anything weird, that’s probably why. Please let us know how you like it!
3
0
23
4,353
Cant wait to see whoever picks Praseodymium first
With Periodic Neon and Gemini Argon now taken, the race is afoot for Krypton
Congrats, Google and the team, on the impressive model and rejoining the frontier! The ongoing intelligence explosion is creating massive consumer surplus for users and startups. PhD on demand, pennies to run.
0
0
5
1,860
Really excited about this model. Well-rounded and incredibly strong across knowledge work, coding, science, and cybersecurity defense.
Introducing Gemini 4 Argon – our new frontier model.
It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
1
0
59
2,967