Signal, not noise.
Releases, benchmarks, and analysis from a curated list of labs, newsletters, and community feeds. Updated every four hours.
Claude Opus 4.7 ships — 1M context, 72% SWE-Bench, new world record
Anthropic’s latest flagship arrives with a 1M-token context window and the highest SWE-Bench score ever recorded on a public model.
DeepSeek V4 quietly reset the cost curve again
At $0.14 input / $0.28 output per million tokens, DeepSeek V4 now matches GPT-5 Mini on quality while costing less than Gemini Flash.
Google ships Gemini 2.5 Flash with 1M context at $0.30/M
Google’s latest cheap-tier model offers a 1M context window and full multimodal inputs at aggressive pricing.
The MCP ecosystem is growing faster than the GPT Store ever did
A year after launch, there are more than 1,200 open-source MCP servers on GitHub. The protocol is winning where GPT actions stalled.
LMSYS Arena drops live leaderboard for three months of review
Citing ongoing research into gameability, LMSYS is pausing live ELO updates through July pending a methodology refresh.
Cursor 1.0 ships with a rewritten agent engine
After three years as a VS Code fork, Cursor hits 1.0 with the most ambitious agent rewrite of the year.
Why we think Llama 4 Scout is the best sub-8B deployment today
Our analysis of price, quality, and fit for self-hosted workloads: why Scout wins for production sub-8B deployments.