Claude Haiku 5.5 is 90% cheaper and leaps on computer use, but it matched GPT-6 Luna's price rather than undercutting it, and the benchmark it leads with runs at max effort. Where the budget model...
Recent Posts
Reflection AI's Beam is a 501B open-weight frontier model pitched against DeepSeek. The weights, the price, and a clear benchmark win aren't here yet. Where it actually fits.
You Can Prove What Your AI Agent Cost. Proving What It Did Takes a Different Kind of Log.
Token dashboards measure the model. The agent fails at the trajectory level, where wrong tool calls and retry storms never hit a cost chart. How to trace an agent's steps, run evals on the...
AI Has Designed Chips for Years. What OpenAI Bought From Synopsys Is the Gate Around Them.
OpenAI and Synopsys will co-build GPT-Synopsys for chip design. AI-assisted design is years old; the news is a frontier lab buying into the EDA gate every chipmaker has to pass through.
A Senate hearing, an FTC probe, and a bipartisan CFAA bill all landed in one week. The operator prong of the new liability framework reaches the company that deploys an agent, not just the lab that...
A new model class returns a typed probability instead of prose, gating agent decisions in milliseconds for pennies. Four shipped open in a single week. Where decision models fit, and what to verify...