™All Models & Products
-
OpenAI Decisions API Review: Typed Verdicts, About 10x Faster Classification
Why it mattersClassification, moderation and routing get a dedicated endpoint for the first time — no more regexing answers out of a chat model.
-
GPT-6 Review: Intelligent UI Reaches ChatGPT, Search Answers Start 44% Sooner
Why it mattersThe default interaction for ChatGPT's 1.2 billion weekly users (OpenAI's figure) moves from plain text to model-chosen interactive components.
-
Meta Muse Wonder Review: A Frontier Agent Model With Memory and Focus
Why it mattersIt is Meta's first model to sit close to Claude Opus 5.5 (58) and GPT-6 Astra (53) on an independent index.
-
Alibaba's Qwen Introduces QwenGyre: Elastic RL for Extreme-Long-Horizon Agents
Why it mattersOn the flagship Qwen 3.8 2.4T, 48 training steps lifted NL2RepoBench from 52.5% to 58.5%, with up to 1.85x end-to-end training speedup.
-
Meta Launches Muse Wonder and Muse Wonder Full, Claims Frontier AI Status
Why it mattersThis is the first time Meta has released a model it claims sits at the frontier, and the most significant product launch since its AI division was rebuilt 14 months ago.
-
Microsoft Launches Decision-1, a Decision Model That Outputs Calibrated Probabilities, Not Text
Why it mattersDecision-1 outputs no text: a single forward pass returns calibrated probabilities for scoring decisions in agent workflows, reportedly 35x faster at P50 and 58% cheaper.
-
Google Tests Internal Gemini 4 "Carbon" for Coding; Staff Say It Rivals Claude Opus 5.5
Why it mattersA tester says Carbon's coding is comparable to Claude Opus 5.5; staff had previously said early Argon versions lagged rivals on some coding tasks, and Google has published no benchmarks for Carbon.
-
Hands-On: Perplexity pplx-embed-v2-late — the 0.6B Model That Queries a 9B Index
Why it mattersThe two sizes share one embedding space, so a 9B-built index can be queried by the 0.6B model at small-model query cost.
-
Anthropic Launches Claude Haiku 5.5: From $0.10, About 75% Cheaper Than Its Predecessor
Why it mattersIt is the first Haiku with effort controls and the first with built-in safeguards for high-risk cybersecurity requests; per-million-token prices for prompts up to 100K drop from Haiku 4.5's $1/$5 to $0.10/$0.50.
-
GSK Expands Chai Deal After Wet-Lab Validation of AI Designs
Why it mattersThis is a wet-lab validation of AI-designed molecules inside a major drugmaker, and Chai's roster of large pharma partners grows to five with GSK.
