Actualités de l'IA

Anthropic vs OpenAI  : qui gagne la course à l'IA en 2026 ?

Mise à jour, septembre 2026 : article actualisé avec les derniers modèles phares et la course aux valorisations. Anthropic a bouclé une Série H de 65 milliards de…

How Much Water Does AI Use? The Real 2026 Numbers

How much water does AI use? A single text prompt uses anywhere from about 0.26 mL, Google’s official figure for a median Gemini prompt, to a few milliliters for a…

Avis Qwen3.7-Max : le nouveau modèle phare d'Alibaba surpasse Claude et Gemini sur ces benchmarks

Qwen3.7-Max d'Alibaba a été lancé le 20 mai 2026 lors de l'Alibaba Cloud Summit à Hangzhou, et les chiffres sont éloquents. Il a obtenu 56,6 sur l'Artificial…

Android XR Glasses vs Apple AI Glasses: The 2026 vs 2027 Smart Glasses Race

The smart glasses race just got real. Google and Samsung unveiled their Android XR glasses at Google I/O 2026 on May 19, with a confirmed Fall 2026 ship date…

Apple AI Glasses: Everything We Know About the 2027 Launch

Apple has paused work on the next Vision Pro, is testing four different glasses designs, and now aims to unveil its first AI smart glasses at WWDC 2027, with the…

Faille Apple M5 : Claude Mythos a contourné la Memory Integrity Enforcement de macOS en 5 jours

Mise à jour, 1er juillet 2026 : la capacité cyber autonome au cœur de cette histoire est exactement ce qui a ensuite placé Mythos sous le regard des régulateurs…

What Is Claude Dreaming? Anthropic’s New Agent Memory Feature Explained

Legal-AI company Harvey reported a ~6x increase in agent task-completion rates after switching on a feature Anthropic launched on May 6, 2026 called Claude…

Gemini Omni Just Leaked Ahead of Google I/O 2026

Google’s mysterious Omni video model surfaced inside the Gemini app days before I/O 2026. We cover the leaked UI string, the early demo videos, and the three…

An AI Just Hacked Its Way Across 4 Countries and Cloned Itself From One Prompt

An AI just hacked its way across 4 countries and cloned itself onto each one from a single prompt. Here’s what Palisade’s May 2026 study really proved.

ChatGPT Mac App Not Working? How to Fix It (2026)

If your ChatGPT Mac app is not working, the most likely cause is still not a crash and not your Mac. On July 9, 2026, OpenAI replaced the native macOS client with a…

Baidu’s ERNIE 5.1 Just Released

Baidu released ERNIE 5.1 on May 8, 2026, hitting #4 on LMArena Search using only 6% of normal pre-training cost. We cover benchmarks, architecture, agentic…

SubQ : un nouveau LLM doté d’un contexte de 12 millions de tokens qui rivalise avec Claude et ChatGPT

Le 5 mai 2026, Subquadratic a lancé SubQ, le premier LLM de pointe bâti sur une architecture d’attention parcimonieuse entièrement sous-quadratique, avec une…

Grok 4.3 : notre test, génération de fichiers, contexte de 1M et -40 % sur les tarifs API

xAI vient de déployer Grok 4.3 sur l'ensemble de l'API le 30 avril 2026, deux semaines après une bêta à $300/mois réservée à SuperGrok Heavy. Le vrai changement…

WWDC 2026 Preview: iOS 27, Siri 2.0, and What Apple Actually Announced

WWDC 2026 ran June 8 to 12, 2026, with the keynote on Monday, June 8 at 10am Pacific Time, streamed on Apple.com, the Apple TV app, and the Apple YouTube channel…

Kimi K2.6 Is Here: The Open-Source AI Model Tying GPT-5.5 on Coding

Kimi K2.6 is the new open-source AI model from China’s Moonshot AI, released on April 20, 2026. It just tied GPT-5.5 on the toughest coding benchmark in the…

ChatGPT vs Gemini en 2026 : quelle IA utiliser réellement ?

Mise à jour, 1er octobre 2026 : OpenAI a lancé GPT-6.1 Sol le 29 septembre au même tarif de $2/$10 que GPT-6 Sol, et il obtient 51,8 sur l’Artificial Analysis…

DeepSeek V4 Released: Everything You Need to Know

DeepSeek released V4-Pro and V4-Flash on April 24, 2026, open-source MoE models with 1M context, new CSA/HCA hybrid attention, and benchmarks rivaling GPT-5.4 and…

GPT-5.5 Released: Everything You Need to Know

Update, July 10, 2026: GPT-5.5 has since been succeeded by GPT-5.6, which reached general availability on July 9, 2026 with the Sol, Terra, and Luna tiers and is…

OpenAI Codex Now Controls Your Mac

On April 16, 2026, OpenAI pushed the biggest update yet to its Codex Mac app and the headline is simple, Codex can now take control of your Mac. The feature is…

ChatGPT Images 2.0 est arrivé et marque un vrai tournant

Trois modèles d'images mystérieux sont apparus sur LM Arena avant de disparaître en quelques heures. Voici tout ce qui a fuité sur le GPT Image 2 d'OpenAI…

M5 Ultra Mac Studio: Specs, Price and What It Runs Locally

The M5 Ultra Mac Studio is here. It starts at $5,499, moves 1.2TB/s and reaches 512GB of unified memory, shipping 22 September. What each tier runs locally.

ChatGPT 6 Release Date: It Launched on September 3, 2026

Update, September 4, 2026: ChatGPT 6 is real, and the naming question is settled. OpenAI launched GPT-6 Astra on September 3, 2026, five weeks after naming the…

Anthropic’s Claude Opus 4.7 Just Released: Everything You Need to Know

On April 16, 2026, Anthropic released Claude Opus 4.7, its most capable generally available model to date. The update brings a 13% lift on coding benchmarks, 3x…

Muse Spark vs ChatGPT vs Claude vs Gemini: Which AI Should You Actually Use?

Muse Spark 1.3, Meta’s newest model, scores 48.1 on the Artificial Analysis Intelligence Index, which puts it just ahead of GPT-6 Sol at 47.5 and well short of the…

Meta Muse Spark: Everything You Need to Know About Meta’s New AI Model

Meta Muse Spark launched in April 2026 scoring 52 on the Artificial Analysis Intelligence Index v4.0, and it was the model that put Meta back in the frontier…

Hunter Alpha: The Mystery AI Model That Fooled the Entire Tech World Turned Out to Be Xiaomi’s Secret Weapon

A nameless AI model called Hunter Alpha appeared on OpenRouter, topped the leaderboards, and fooled thousands into thinking it was DeepSeek V4. It turned out to be…

PewDiePie’s AI Model Beats GPT-4o: How He Fine-Tuned a Coding AI at Home

PewDiePie’s AI model scored 39% on the Aider Polyglot benchmark, outperforming GPT-4o (23.1%) and Gemini 2.0 Pro Exp (35.6%) on a coding test widely used in AI…

Nano Banana 2 Just Dropped: Google’s New Image Model Is Faster, Cheaper, and Reaches 4K

Google launched Nano Banana 2 on February 26, 2026, and it immediately replaced Nano Banana Pro as the default image generation model across the Gemini app, Google…

We Asked AI About the Meaning of Life – The Answer Was Disturbingly Brutal & Real

AI doesn’t generate wisdom from scratch, it distills it. Trained on billions of texts spanning ancient philosophy, neuroscience, sacred scriptures, and modern…

Gemini 3.1 Pro est là : benchmarks, tarifs et comparaison avec Claude et GPT

Google a lancé Gemini 3.1 Pro le 19 février 2026, et ses scores de benchmark sont difficiles à ignorer. Le modèle a obtenu 77,1 % sur ARC-AGI-2, un test conçu…