Fable 5 landed with an 80.3% SWE-Bench Pro score. Gemini 3.5 Pro missed its own June window again. Grok 5 is still training. Here is what developers can actually use today.
AMD's EPYC Venice becomes the first HPC chip on TSMC N2, packing 256 Zen 6 cores and claiming over 70% performance gain over Turin with 1.6 TB/s per-socket memory bandwidth.
Google's Gemini 3.1 Flash-Lite lands at $0.25/M input tokens with a 1M context window. Here's which production workloads justify switching - and which don't.
On the same day in May 2026, OpenAI made GPT-5.5 Instant the ChatGPT default while startup Subquadratic launched SubQ, a 12M-token model that ditches the transformer's quadratic attention. Two opposite bets on AI's future.