OpenAI's new GPT-6 Sol and Luna models offer lower API prices, and updated prompt caching could cut costs further for apps that repeat the same instructions. Anthropic has a faster Claude Opus 5.5, while Meta's Muse raises a practical question: what access should we give an AI assistant?
Read today’s briefing Curated AI News and Tool Recommendations
What’s worth knowing and using in AI.
We curate the model drops, try the latest tools, follow the cutting-edge research, then distill it all just for you.
Latest briefing · 2026-09-23
GPT-6 gets cheaper, and Claude cuts costs too
01
The Big Story
Introducing GPT-6 Sol and Luna
02
For Builders
Better prompt caching for GPT-6
03
For Business
Meta admits Muse’s likeness to OpenClaw isn’t a coincidence
Tested, researched, or worth watching
Tools that we recommend (and use ourselves ).
Hands-on harness
Pi Agent
Pi is a minimal terminal coding harness that can be adapted with extensions, skills, prompts, and themes.
Researched product
Claude Cowork
Claude Cowork works across the files, tools, and browser sessions a user selects, then brings the finished work back for review.
Hands-on platform
LM Studio
LM Studio provides a desktop application and local API for downloading, running, and building with open models.
Start with your needs
Explore by use case.
Coding
Agents and developer tools that can produce work we can verify.
Research
Search and synthesis tools that keep the original sources close at hand.
Productivity
Tools that reduce routine work without adding another system to babysit.
Creative
Image, video, audio, and design tools that are ready for real projects.
Business
AI for operations, customer work, sales, and better decisions.
Models
The models and infrastructure that matter beyond a benchmark score.
Analysis and guides
Deeper coverage on how to use AI effectively
Where coding agents are useful today
The work coding agents handle well, the tasks they still get wrong, and where experienced developers need to step in.
How to choose an AI research assistant
A practical way to compare citation-first search, deep research, and general-purpose assistants with web access.