Aller au contenu principal

AI Models & Assistants

Tous les niveaux

Claude is our core — but the same skills travel. This section widens the lens to the whole AI world: the major assistants, how they differ, when to use which, and the techniques that transfer across all of them.

What you'll learn
  • Pick the right model for a job without re-learning the field every time
  • Move between assistants — ChatGPT, Gemini, Grok, open models — without losing your technique
  • Run capable models on your own machine, and know when that is worth it
  • Combine Claude with local models and agent frameworks

Start here

If you read one page in this section, read this one. Everything below is a branch off it.

Then follow whichever thread matches what you are doing.

The current flagships

The two September 2026 releases at the top of both stacks, read side by side.

Coming from another assistant

You already know one tool and want your habits to carry over.

Comparing before you commit

Running models yourself

Local models trade some capability for privacy, cost control, and working offline. These pages cover when that trade is worth making.

Agents beyond Claude

Everything in this section

The full list, including pages added since this index was written.

Mise à jour GPT-5.6 d'août 2026 : curseur d'effort, bouton Think gratuit et le précipice tarifaire des 272K

Le 6 août 2026, OpenAI a livré une mise à jour de mi-cycle pour GPT-5.6 que la plupart des articles ont résumée à « Sol est devenu plus intelligent ». La vraie histoire : un contrôle d'effort à six niveaux qui change la façon dont vous tarifez une requête, un bouton Think pour le niveau gratuit, un précipice tarifaire à 272K tokens qui re-tarifie tout l'appel, et un mode Fast qui ne fonctionne que sur Sol. Ce que chacun signifie pour un utilisateur de Claude.

Apple Foundation Models 3 & la stack LLM on-device pour utilisateurs de Claude

La troisième génération des Foundation Models d'Apple (AFM 3) a atterri à la WWDC 2026 avec un modèle dense on-device 3B, un MoE sparse 20B qui tourne sur iPhone, et un framework Swift-natif qui vous permet maintenant d'apporter votre propre fournisseur LLM — y compris Claude. Ce qui a réellement été livré, les mécanismes non évidents, et comment ça change l'histoire on-device pour quiconque livre des apps Claude.

Navigateurs à session partagée pour agents IA : le modèle ego lite

Une nouvelle catégorie de navigateur — parfaitement illustrée par ego lite, qui a atteint la #1 place du GitHub Trending le 24 juillet 2026 — permet à Claude Code, Codex, Cursor et à tout autre agent de coding de piloter une session Chromium isolée qui hérite de vos véritables connexions. Ce que les Spaces et les snapshots sémantiques résolvent réellement, pourquoi ils surpassent Playwright pour le travail d'agent, et le coût de sécurité lié au partage des cookies de session avec un processus autonome.

Cloudflare Kitesurf: The First Browser Runtime Built for AI Agents (Not Humans)

On August 6, 2026 Cloudflare launched Kitesurf, a stateless, agent-first browser that runs entirely in V8 isolates on Workers. It uses 3-7x less CPU and memory than Chromium on common agent tasks — at the cost of ~1.7x wall-clock time. Written in Rust, using Firefox's CSS parser (Stylo) and the Blitz rendering engine, drop-in compatible with Puppeteer/Playwright via the Chrome DevTools Protocol. This page: what changed, when to reach for it instead of headless Chrome, how to wire it into Claude Code / Codex / your MCP client, and the four things it flat-out can't do yet.

Colibrì: Run a 744B MoE From Your SSD (and What It Really Costs You)

Colibrì is a pure-C, zero-dependency engine that runs GLM-5.2 (744B), Kimi K3 (2.8T), DeepSeek V4.1 Flash and six other frontier MoE models on a 16–32 GB machine by keeping the dense layers in RAM and streaming the routed experts from NVMe. This page explains the memory hierarchy, the numbers that decide whether it is usable for you (0.05 to 6.8 tok/s), the SSD-wear and speculative-decoding gotchas, and the exact commands to get a first token.

OpenAI Agents API vs Claude Managed Agents: The Hosted Agent Loop, Compared

On September 10, 2026 OpenAI put the Codex harness behind a single API call: the Agents API public beta — durable sessions, hosted or self-hosted sandboxes, automatic compaction, tool search, subagents. Anthropic has run the same shape since April as Managed Agents. Both use the same four primitives; they differ on permissions, secrets, budgets, network policy and what the sandbox costs. The side-by-side, the request shapes, the six gotchas, and when to pick which.

Next