Free users get unrestricted text chats and a Think button for harder questions; Plus/Pro get a sharper Sol model that unifies Instant and Thinking into one slider.
Why it matters: Unlimited free-tier access lowers the bar for marketing teams to pilot frontier models — fewer seat-purchase decisions blocking proof-of-concept starts.
The API alias silently switched on August 5; developers staying on 1.0 needed to pin before the migration, while 2.0 delivers better transcription and tool-calling accuracy at $0.08/min.
Why it matters: Voice agents just got cheaper and more accurate from a major provider — a useful benchmark when evaluating speech-to-speech features for automation workflows.
Team and Enterprise orgs can now run sessions on their own runners (fixed or on-demand scaling); agents can message each other across sessions via permission-gated SendMessage.
Why it matters: Operators with data residency or compliance constraints can now run Claude Code on their own infrastructure, and multi-agent workflows gain a direct session-coordination layer.