Introducing GPT-5.2 🚀
🧠 Expert-level knowledge work — wins or ties pros on 70.9% of GDPval tasks
💻 SOTA coding — 55.6% on SWE-Bench Pro and 80.0% on SWE-bench Verified
🖼 Sharper vision — better at charts + UI understanding
🧩 Long-horizon tool use — 98.7% on Tau2-bench Telecom
📚 Long-context reasoning — stronger performance on MRCRv2 up to 256k tokens
#frontend #developer #react
Instagram ⏺ YouTube ⏺ Threads ⏺ Medium ⏺ X
Post #1078
202
