Tim B.

@RejaullahmdMd @ai_for_success Fake news

NEET 2026 was conducted yesterday in India. Within 24 hours I tested 8 frontier AI models on all 180 questions. 🥇 Gemini 3 Flash: 715/720 (99.3%) — would rank AIR-1 🥈 GPT-5.5 & Gemini 3.1 Pro: 710/720 (98.6%) Full results 👇

@thsottiaux It should be a separate product please don't change chatgpt. Its okay when it use the same model

What are we obviously not getting right with Codex?

@l_mejiaC @sama Where?

Sam any hints on AVM v2 @sama 👀

@l_mejiaC Chatgpt 5.5 is much better than 4.6. 4.5==5.5.

Opus 4.6 is still the best model to chat with. It has presence & is really good at explaining things. And no, 5.5 is not the new 4.5, it still needs a soul. A souler soul. 5.5 is considerably smarter though

@Angaisb_ @immasiddx @mark_k Codex will never replace CHATGPT App! Chatgpt is the main product

@immasiddx They'll probably end up replacing the ChatGPT app with Codex @mark_k said this and Tibo basically confirmed it

Claude’s macOS app looks so much better than ChatGPT. It’s not even close. 😭

@immasiddx Fake news

@haider1 Fake news

remember when gpt-4.5 was supposed to be many years away from running on consumer hardware? yet 12 months later we're running something several generations better on our laptops with qwen / gemma models the architecture keeps improving in ways we can't fully predict

@0x_Vivek The middle May! 2 Month difference to much!

openai model release cycle is clearly compressing: gpt-5: aug 7, 2025 gpt-5.1: nov 12, 2025 — 97 days later gpt-5.2: dec 11, 2025 — 29 days later gpt-5.3-codex: feb 5, 2026 — 56 days later gpt-5.4: mar 5, 2026 — 28 days later gpt-5.5: apr 23, 2026 — 49 days later gpt-5.6 should launch around early june

@daniel_mac8 Mythos is a Mythos it doesn't exist only marketing

Claude Mythos this week? Last few Anthropic dev conferences brought new models. Based on the read of dev vibes, Anthropic may need to make Mythos GA sooner than later.

@daniel_mac8 Mythos will never release

@l_mejiaC Then all people don't have a job anymore

I’m ready for super intelligence.

@Angaisb_ Already available in Codex

OpenAI could make a similar thing to Claude Design where GPT-5.5 generates websites in different styles using GPT Image 2 as reference and then the users choose one of 2-4 options They'd get a ton of UI taste data with that

@AlexMasmej Nicht viel? 100$ in one day is very big

I'm using Codex 10x more than Claude, and a reason why is despite being on the business plan I constantly hit rate limits. I'm not using it that much and bought $100 of credits yesterday, already gone.

@haider1 Fake Benchmarks

grok 4.3 took the top spot in instruction following and i think it doesn't surprise me i pick grok especially for non-agentic tasks because, since grok 4.1, it has become much better at following instructions accurately and consistently it sticks to what you ask without adding extra stuff or skipping parts of the request

@BlockView0214 @swyx @xai @grok Fake news

@swyx @xai @grok Codex is quietly turning from a coding tool into a new ChatGPT entry point. Chat is great for asking. Codex is better for working: files, shell, repo context, task state. The next AI app battleground is not “who chats better”, but who owns the workspace.

small milestone: uninstalled the chatgpt app. codex is strict superset now! found something cool - among frontier models, @xai @grok 4.30 is the most intelligence per dollar you can get, beating even open models like MiMo, Kimi, and DeepSeek. numbers pulled from @ArtificialAnlys

@mark_k @OpenAI Fake News

ChatGPT is going to merge into the Codex app! 🔥 This was confirmed by an engineer of Codex at @OpenAI. The combined “super app” will bring together the more casual, general-purpose ChatGPT experience with the coding-focused power of Codex. In other words, one app for chatting, reasoning, writing, coding, debugging, and shipping. It makes a lot of sense. ChatGPT is becoming more agentic, Codex is becoming more general-purpose, and the line between “AI assistant” and “AI developer environment” is getting blurrier every week.

@mark_k It never happen

Now that OpenAI ChatGPT is going to merge with Codex, how should the fusion be named?

@haider1 4.5 is integrated into 5.5

gpt-5.5 is good, but gpt-4.5 was a tasteful beast we haven't seen a model that big with reasoning abilities because it would be too expensive to serve since gpt-4.5 was released, RL has made same-size models far smarter add reasoning to that, and things change completely

Even though 5.5 is a very good model We still have a medium way to go till AGI and I am excited for 5.6, 5.7 & GPT 6 and what the frontier will look like at the end of the year.

@nicdunz It's not possible 🤣

there hasnt been a chatgpt ios app update in like a year