Today in AI – May 22, 2026

May 22, 2026 — curated links and takeaways.

1. OpenAI Introduces GPT-5.5

OpenAI released GPT-5.5 on April 23, 2026, positioning it as their smartest model yet with improvements in coding, research, and data analysis across tools. Developers need benchmarks against competing models to assess migration value and capability gaps.

2. Google Releases Gemma 4 Open-Weight Family

Google released Gemma 4 on May 4, 2026, expanding the open-source reasoning model ecosystem. Open-weight models lower deployment costs and reduce vendor lock-in for teams building production systems.

3. Anthropic Doubles Claude Code Rate Limits, Raises API Caps

Anthropic doubled rate limits on Claude Code and increased API limits for Claude Opus to enable developers to build reliably at scale. Infrastructure constraints are critical for production deployments—these changes directly impact throughput for startups and enterprises.

4. Anthropic Launches Project Glasswing with $100M Model Credits

Anthropic committed $100M in model usage credits to Project Glasswing, offering Claude Mythos Preview at $25/$125 per million input/output tokens. This security-focused initiative signals consolidation around enterprise AI safety standards.

5. SpaceX Acquires xAI as Subsidiary, Allocates $7.72B to AI Infrastructure

SpaceX acquired xAI in February 2026 with plans to expand Colossus to 1M+ GPUs, spending $7.72B on AI in Q1 alone ahead of its Nasdaq IPO. This consolidation signals major capital deployment in compute infrastructure and represents a structural shift in AI resource concentration.