Labeled default for coding (~73% ideal), research (~59%), and creative (~56%). Strongest overall prior in founder history.
Measured math/logic specialist (Claude-family ~59% ideal; pairwise harness preferred Opus). Not the coding or creative default.
03 · Google
Gemini 3.1 Pro
Present in labeled history as a secondary pick; no category default yet. Kept for coverage, not as a proven specialist.
General-chat default: 86.5% good when Grok was actually deployed (45/52). Not the diluted 41% ideal-share figure.
Current-info specialist from labeled history (forced when needs_current_info). Strong second on research ideals.
Live web-search option with thin labeled support (only 5 current-info ideals). Available, not the current-info default.