Tracked topic
… Sign in to your XDA account Summary OpenAI launched GPT-5.5 just weeks after GPT-5.4 — rapid version churn continues. GPT-5.5 posts benchmark gains, notably +7% on Thinking with Terminal-Bench 2.0. Available now in ChatGPT and Codex for Plus, Pro, Business, and Enterprise users. …
… I'm sure GPT-4.5 has strengths that I'm not utilizing, but if I'm not using them why would I use the model I don't like using? Plus, I can't stand the fawning nature of GPT There is one reason to use GPT-5.4: the larger context window. …
… Now, ChatGPT has revealed its own takes on Anthropic's releases, with GPT‑5.4‑Cyber as the model and an expanded Trusted Access for Cyber TAC as the framework. Related I cancelled my ChatGPT, Perplexity, and Gemini subscriptions for Claude — and I should have sooner Wish I did this sooner. …
… Over time, I started to treat these models less like presets and more like sub-tools; for writing codes, I used GPT 5.3-codex, GPT 5.3-mini for fast answers, and GPT 5.4 for general question answers. …
… Related I used Claude, Gemini, and ChatGPT to create my personal website, and there was a clear winner The AI web dev battle. ChatGPT decided to take a rather generic approach It's nothing I haven't seen before GPT-5.5 took my resume and immediately began telling me what was wrong with it. …
… GPT-5.5 scored 82.7% on Terminal-Bench 2.0, blowing past Opus 4.7's 69.4%. …
… Sam Altman described the model as "going from GPT-3 to GPT-5" all at once, which is a fairly bold claim to make. That said, ChatGPT's initial image generation model is something that I and a lot of other people found pretty underwhelming. …
… ChatGPT faltered, but not by a lot It didn't match up to Claude, though By this time, I knew exactly what to ask ChatGPT to create for me, and OpenAI's GPT-5.5 needed to outdo Sonnet 4.6, or at the very least, compete toe-to-toe. …
… ChatGPT was the clear and obvious front-runner for general use when I began using one daily, so when it became time to purchase a subscription, ChatGPT Plus was what I went with. …
… It's a 230B mixture-of-experts with only 10B active parameters per token, eight of 256 experts routed per token, a 205K context window, and a near-GPT-5.3-Codex score of 56.22% on SWE-Pro. On Terminal Bench 2.0 it scores 57.0%, which is the same ballpark as GPT-5.3-Codex. …