v0.1.3, pricing breadth and example app fixed

Closes the two highest priority adoption blockers from the v0.1.2 review.

Pricing table expanded

Went from 9 entries to 50 plus across the three providers:

• Anthropic: Claude 4 family (opus, sonnet, haiku), Claude 3.7, 3.5, and 3 era
• OpenAI: GPT-4.1 / 4o / 4 turbo / 4, o1 / o3 / o4-mini reasoning models, GPT-3.5-turbo
• Google: Gemini 2.5 / 2.0 / 1.5 family (pro, flash, flash 8b)

Best effort verified as of 2026-05-01. The pricing PR template at .github/PULL_REQUEST_TEMPLATE/pricing-update.md is the path for community corrections.

Example app deps pinned for real

The example at examples/expo-dev-client/ now uses the version pins Expo SDK 53 actually expects, verified via npx expo-doctor and npm run typecheck. Previous pins were guesses that would have produced dependency conflicts on a fresh install.

Tests

171 unit tests, 99.4 percent line coverage. Four new pricing breadth tests catch regressions in model coverage.