Heidelberg AI News

Heidelberg AI News

All stories

4 stories about Genie 4

Clear

Thursday 29 October

  • Latent Space

    ArticleLatent Space

    Can Frontier Models See Past a Decoy? Elorian Tests Visual Reasoning

    Elorian, which emerged from stealth in April with $55 million in seed funding co-led by Striker Venture Partners, Menlo Ventures and Altimeter Capital, launched its first visual reasoning model, Glimpse, described as a 'visual thinking model' that reasons about the physical world from what it sees. In partnership with AI data company Mercor, Elorian co-released DecoyCount, a benchmark testing whether models genuinely count objects or rely on visual shortcuts and substitution heuristics. According to a paper accompanying the release, frontier models scored below 50% accuracy, with GPT-6 Astra highest at 43.17%. CEO Andrew Dai said Glimpse scored 50.48% on 1,660 DecoyCount items in an internal test, making it the only model to clear 50%, though he noted the result came from a single run versus the frontier models' three-run averages. Dai, a former Google DeepMind researcher on GLaM MoE, PaLM 2 and Gemini, said the model uses visual thinking—generating and editing images as part of the reasoning trace—and distinguished Elorian's 'indirect action' approach from world models like World Labs' Atlas and Google's Genie 4.

Wednesday 28 October

  • heise online

    Articleheise online

    Google Genie 4: KI-Simulator für interaktive 3D-Umgebungen wird realistischer

    Googles Genie 4 erzeugt interaktive KI-Welten mit Ton und Sprache. Gemini verbessert die Simulation, doch die Technologie bleibt experimentell.

  • ArticleLatent Space

    [AINews] Anthropic ends Coding Interviews: "Hunger beats tenure"

    Latent Space's AINews episode covers several AI developments. Anthropic hiring manager Felix Reiseberg announced he has stopped asking candidates to do coding interviews, arguing Claude now outperforms people at React and assembly tasks; he advises junior engineers to prioritize business impact over backlog, pursue ambitious ideas, build working prototypes over a weekend, and lean into their fresh user perspective ('Hunger beats tenure'). Google DeepMind introduced Genie 4, a real-time world model with native audio, physics, and Gemini-based planning, positioned toward testing agents and robots. In AI for science, Axiom announced a Lean auto-formalization of the Classification of Finite Simple Groups (~15,000 pages across ~500 papers, ~50M lines of Lean produced by agent swarms in 24 days), displacing the four-color theorem as the largest computer-verified proof, while OpenAI and Anthropic reported independent claims related to disproving the Exponential Time Hypothesis. Independent evals were updated: Artificial Analysis Intelligence Index v5, FrontierCode 2.0, and results for a new GPT-6.1 model.

  • Golem.de

    ArticleGolem.de

    Google Genie 4: Simulator zeigt fotorealistische, per Prompt steuerbare Welten

    Google Deepmind startet Genie 4 als Research-Preview. Das Weltmodell erzeugt Ton und folgt Textbefehlen live. ( Deepmind , KI )