Callstack launches Apex coding model: near-Opus React Native quality at 27% of the cost
aidenybai · x · 2026-10-14
Callstack released Apex, an agentic coding model optimized for React Native and web React.
- Scores 87.8% on React Native evals, within 1 point of Claude Opus 5.5 (88.6%), at $7.72 per task vs $28.75 for Opus 5.5 (27% of the cost), with up to 250 tokens/s throughput (3x GPT 6.1 Sol)
- On Next.js it hits 96.8% Pass@1, matching GPT 6 Astra at 1/13th the cost
- Ranks #2 on fixing React (22%) and tied #5 overall (35%) on aidenybai's ReactBench, showing React Native tuning transfers to the web
- Works with Claude Code, Cursor, and Codex via endpoint swap; available on Vercel AI Gateway and open-sourced at callstack/apex
More from coding & agent
- Claude builds its own test tool, fixes DreamPark MR sky-flicker bug in minutes — nptacek · 2026-10-14
- Argon arrives in Google Antigravity but users say it's useless even on the 5x plan — waqasalibrand · 2026-10-14
- CMU paper: a 4B model rewrites an agent's harness and beats the 35B model it learned from — irinarish · 2026-10-14
- Harness ≠ infrastructure: sandboxing and persistence belong to a different layer — blaizedsouza · 2026-10-14
- Anthropic publishes security guide for locking down Claude Managed Agents sandboxes — brada · 2026-10-14
- Dev used Gemini as software architect to build an Apple Watch caffeine tracker — pythononrailz · 2026-10-14