Shippty ship ship, McShipface.

Trust the process.

We’ll reset usage limits for all paid users across codex and ChatGPT work!

Colombia withdraws from the baseless 'genocide' claim against Israel. There is no evidence of intent, policy, or campaign to destroy in whole or part the people of Gaza. Actually a mountain of evidence contrary to the accusation.

📝 Simple, scannable editing in @ChatGPT When writing in chat, you'll now see blue text to highlight what's been added or changed I've found it makes it so much easier to collaborate with chat and see at a glance what's changed btwn turns Small one ahead of a big next week..

Well said Robin, thank you.

I went on a long ass walk™ with ChatGPT Voice. Over four hours, I worked with ChatGPT using only my voice. We had a longgggg conversation about my whole professional career, from my BNY Mellon days when I first started my career in financial services at 18 through to today, where I'm an AI founder and engineer. Together we collected the stories, the achievements and the bits of my working life that matter. We also got into what brings me joy, what I'm personally interested in and the wider human experience I've had because it all feeds into who I am, what I bring to the table, and how I can help more businesses and clients. What I really appreciated: - I could ask questions, ask it to repeat things and get it to reword what it had shared - It was connected to my apps: my knowledge base, Notion, Google Drive and emails. It pulled that context into the conversation, then asked specific questions and shaped stories around what I'd actually achieved - It drilled into the detail. A lot of it I'd honestly forgotten about, and I was genuinely surprised - I got to step away from the screen and the desk and still be productive. 25,000 steps and four hours later, I have my whole past history captured. What I didn’t enjoy: - it would get to about 30 mins into a convo and cut out, with no audible sound so you lose what you’ve said and have to reconnect and repeat yourself Either way, now it’s saved forever in my second brain (elliott-professional-experience dot md), can become linked case studies and stories that help me grow my client base and put me in the best possible position for the next chapter. Most only use AI sat at a desk, typing into a chat box. But it can listen, ask questions and pull in your own context while you move. Voice is the interface of the future and gives you the freedom to do real work. Have you tried working with AI like this yet? (not sponsored, video edited by Opus 5.5)

computer (Astra) lights on please. this may be the most extensively detailed & accurate 3d scene ever created by an ai. 400+ hours of Astra

GPT-6 Sol is the most cost-efficient model we've ever seen on Vending-Bench.

Claude Opus 5.5 Max actually really sucks compared to GPT-6 Sol Max: -1.8x the cost per correct task -68% more fabrications -well it does complete .4% more finished jobs (LoL) Why? Opus 5.5 requires 3x the output tokens to get to the correct output.

My first test of GPT-6 Astra vs Claude Opus 5.5: a simple Blender task, same prompt, same goal, both at Max. The results look almost the same to me, but Astra was much more efficient. Astra: 43 min, 61K output tokens Opus: 1h 57 min, 507K output tokens Deeper tests next.

GPT 6 Sol beats Fable 5 on BridgeBench. At 1/5 of the cost. GPT 6 Sol: 643 Fable 5: 628 GPT 6 Sol is one of the best models I have ever used. And on a ChatGPT Pro subscription the usage limits are basically unlimited.

Across my 12 experiments Opus 5.5 vs. GPT-6 Astra

To me, this is the original vision birthed with GPT-4o that is now realized. Real time, bi-directional —smart— voice that can directly leverage the entire ecosystem of plugins is a bigger unlock than it appears at first glance. Compounding benefits.

Claude Opus 5.5 is the #2 model in the world at filing taxes according to TaxCalcBench. Unfortunately, it's also 10x more expensive than the new #1 model: GPT-6 Sol.

gpt-6-luna (flex, low thinking) is phenomenal for text extraction. Half the cost of the next cheapest and just as accurate as gpt-5.6-luna. I switched from gemini-2.5-flash-lite when I started getting refusals for text extraction (?). Much better quality with luna in general.

I just had a voice conversation with the Datasette backup of my blog via datasette-mcp, which you can now talk to using the ChatGPT iPhone app

We tested Opus 5.5, GPT-6-Sol, and GPT-6-Luna thoroughly across 100 open-ended coding and engineering environments. Our results have significantly diverged from AAII. Some takeaways: • GPT-6-Astra is still the frontier model by a comfortable margin. • Opus 5.5 ranges from #2-#7 on coding categories, being the second best model in its ability to one-shot code (which is our best correlated measure with fluid intelligence). We don't test "usability," but anecdotally, it seems they improved its communication style. And the lower pricing is a welcome change -- we can probably thank @thsottiaux for that. It's not a cheap model, but it's the first reasonably priced Anthropic model. However, Opus 5.5 underperformed in our custom harness, which is unusual for Claude models, and a pattern that started with Fable 5.1 (one-shot-fluid intelligence measured better than agentic results). One observation is that it frequently self-stopped when it felt its results were good enough: e.g. in one evaluation, Opus 5.5 closed the session, with the reason: "averaged 867,703 in the latest evaluation, ranked 2nd of 31 against sampled opponents"... so I suppose that saves money, but it still has a tendency to make questionable unilateral judgment calls like that. In our harness, almost every other frontier model keeps working until their code stops improving or they use their full call budget. I do wonder how much of this is due to Anthropic retuning their "High" effort mode (which we use for testing)*. • GPT-6-Sol is within margin of error of Opus 5.5, and Pareto optimal*. It's faster than all Anthropic models, even on a Flex endpoint. Competitive with the frontier on intelligence and cost. • GPT-6-Luna is slightly smarter than GPT 5.6 Luna, and discounted. The old Luna already had a lot of applications at its price point, so this model might end up being the most impactful of the group on non-engineering work. About GBENCH from @GertLabs: our environments are open-ended with no single correct answers, but verifiable results. They involve creativity, design, and are often organized as multi-agent coding games or engineering design challenges. You can see some live demo examples at gertlabs.com/spectate. Every one of our environments in the benchmark pool is unsaturated, and we remove any eval that shows any signs of saturation/stagnation across multiple releases. It's designed to measure and differentiate uncontaminated, raw intelligence for frontier models specifically. These results are interesting and we've thoroughly reviewed them. I don't like to see Opus 5 still so high on the leaderboard, but ultimately I think this is more a problem with Anthropic's newest releases, and it probably explains the price decrease and consecutive releases since Astra came out. They haven't significantly improved upon Opus 5's intelligence, only its personality (which is honestly still a huge win tbh). *We don't chase marketing terms like xhigh and max. We test all models on adaptive/auto-reasoning where supported and "high" effort where configurable. Some companies beef up their max effort more than others (making them unrealistically slow and expensive in practice), but "high" shows you what an underlying model is capable of and is the more common configuration. This is likely working against Opus 5.5 here. *Some caveats on price and speed -- we have been using the "OpenAI Flex" endpoint since the Astra release, which I recommend you try if you're using API pricing. It's half price and a little slower, but OpenAI models are already quite fast so it hasn't been an issue in practice. But keep that in mind when making price comparisons to earlier OpenAI models on our cost and speed efficiency charts.

ChatGPT Finances now supports @Coinbase connections through @Plaid! This has been one of our most requested features. Thank you for your patience.