Everything happening in AI· Tracking 200+ sources so you don't have to
This Week's Signal
The biggest AI news from the last 7 days
Claude Opus 4.6 — 1M context, agent teams
Features a 1-million-token context window and introduces agent teams — groups of agents that can split larger tasks into segmented jobs. Outperformed GPT-5.2 on several benchmarks.
Claude Sonnet 4.5 — autonomous agent benchmark leader
State-of-the-art on SWE-bench Verified and OSWorld (61.4%). Can work autonomously for 30+ hours (up from 7 hours with Op...
Claude Sonnet 4 and Opus 4 launch
Claude Sonnet 4 delivers superior coding and reasoning as a hybrid model. Claude Opus 4 is declared the world's best cod...
o3 — OpenAI's smartest reasoning model with tool use
OpenAI releases o3, its most capable reasoning model, making 20% fewer major errors than o1. The first reasoning model t...
Llama 4 Scout and Maverick — MoE + 10M context
First Llama models with native multimodality and mixture-of-experts architecture. Scout supports a 10M-token context win...
Gemini 2.5 Pro — Google's first thinking model
Google's first 'thinking model' that reasons through steps before responding. Debuted at #1 on LMArena by a significant...
Claude 3.7 Sonnet — first hybrid reasoning model
The first hybrid reasoning model on the market, able to produce near-instant responses or engage in visible extended ste...
Grok 3 — trained on 200K-GPU Colossus supercluster
Trained with 10x the compute of Grok-2 on the 200,000-GPU Colossus supercluster. Features a 1M-token context window and...
o3-mini — fast, affordable reasoning
OpenAI releases o3-mini, a cost-efficient reasoning model optimized for STEM. Excels in science, math, and coding, showi...
DeepSeek R1 — open-source reasoning model shakes markets
An open-source reasoning model rivaling OpenAI o1. Its mobile app briefly surpassed ChatGPT as #1 free app on the US App...
Trending Models
The AI models making the most noise right now
Claude Sonnet 4.5
Anthropic's best balance of speed and intelligence. Strong coding and reasoning at an acce...
GPT-4o
OpenAI's fast multimodal model. Natively processes text, audio, image, and video inputs. G...
DeepSeek V3
DeepSeek's 671B MoE model (37B active) offering near-frontier performance at extremely low...
Gemini 2.0 Flash
Google's fastest model with native tool use, multimodal output, and agentic capabilities....
Quick Take
What's shaping AI this week in three lines
The race is on
Top AI labs are competing harder than ever to build the smartest model
Free AI is catching up
Open-source models are getting close to matching the paid ones
AI that works alone
Autonomous AI agents are going from experiment to everyday tool
Explore
Go deeper into each area