Anthropic has adjusted the usage limits for Claude Code, its AI-powered coding agent, in a change that on the surface appears to increase capacity but effectively reduces usable work according to posts on X and Anthropic's developer forum.
What Happened
Claude Code offers different tiers with message and turn limits per pricing period. Anthropic increased certain numerical limits, according to The Decoder's reporting, which characterized the shift as nominally an improvement that effectively reduces usable capacity due to how turns are counted and reset. Developers posting on X and Anthropic's forum described the change as resulting in less work getting done despite the higher stated limits.
Why It Matters
For developers relying on Claude Code for sustained coding tasks, a limit framed as higher can still mean fewer completed operations if accounting methods change. This affects workflows where long-running sessions or large refactors require consistent token and turn availability. Pricing tier comparisons become harder when the definition of what counts toward a limit shifts without corresponding increases in practical throughput.
The Bottom Line
Anthropic's update to Claude Code's limits illustrates how headline numbers can mask substantive changes in usable capacity. Developers evaluating the tool should examine current usage patterns against the new counting methodology before assuming increased headroom.