The following three changes—all effective today—are aimed at improving the experience of using Claude for our most dedicated customers.
First, we’re doubling Claude Code’s five-hour rate limits for Pro, Max, Team, and seat-based Enterprise plans.
Second, we’re removing the peak hours limit reduction on Claude Code for Pro and Max accounts.
Third, we’re raising our API rate limits considerably for Claude Opus models, as shown in the table below:
A recurring problem with generative AI and Prolog code is that models frequently use =/2 instead of proper comparison operators. This issue has persisted for years and continues to occur even with recent Anthropic models like Haiku 4.5 and Sonnet 4.6. (Note: I typically reserve Opus 4.7 for planning rather than coding tasks.)
When I asked Haiku 4.5 to explain this behavior and provide a reference, it cited:
In other words, and unless you are writing hairy control flow code, if you are using “==” you are probably doing it wrong. Use “=” instead.
The model attributed this to the official SWI-Prolog reference manual. However, I was skeptical—I couldn’t imagine @jan writing such advice as a documented fact—so I traced the source. While the quote does appear on the SWI-Prolog page for Comparison and Unification of Terms, it’s actually a user comment at the end of the page, not official documentation.
Anthropic is also recommending a limit to the number of skills to have:
