Image created with Ideogram V2. Image prompt: A vibrant spring meadow with exaggerated blooming flowers in bright colors. Hidden comically in the middle is the OpenAI logo structure attempting to blend in with flowers growing from its geometric shape. ChatGPT interface windows pop up between flowers. DALL-E easels with AI-generated art sit partially concealed. Woodland animals wear tiny lab coats conducting experiments. A paper with “GPT-4” written on it flutters in the breeze. The whole scene is bathed in golden sunshine with lens flares. Vibrant colors and high detail. The word “OPENAI” integrated into the scene.
OpenAI Developers on X: “Meet Codex CLI—an open-source local coding agent that turns natural language into working code. Tell Codex CLI what to build, fix, or explain, then watch it bring your ideas to life. https://t.co/jjPZdRIgrm” / X
https://x.com/OpenAIDevs/status/1912556874211422572
Rowan Cheung on X: “Some other key updates: Both new models are able to use all ChatGPT tools independently (web browsing, Python, images, etc.). Available now for Plus, Pro, and Team users, and will be replacing o1, o3-mini, and o3-mini-high. o3-pro is coming in “a few weeks”. https://t.co/risspxyDUL” / X
https://x.com/rowancheung/status/1912561389824070120
Rowan Cheung on X: “CFO Sarah Friar also recently said that o3-mini is already the No. 1 competitive coder in the world: https://t.co/wsHPQqKI5I” / X
https://x.com/rowancheung/status/1912561394068697210
Rowan Cheung on X: “The open-source Codex CLI agent launching today: — Runs locally in terminal — Interface to “link models with local code and computing tasks” — Built for o3 + o4-mini, GPT-4.1 support coming soon — $25K API credit grants are available for early projects https://t.co/p71wVqdsto” / X
https://x.com/rowancheung/status/1912561395591241819
“o3 and o4-mini are super good at coding, so we are releasing a new product, Codex CLI, to make them easier to use. this is a coding agent that runs on your computer. it is fully open source and available today; we expect it to rapidly improve.” / X https://x.com/sama/status/1912558495997784441
Rowan Cheung on X: “The biggest change: o3 and o4-mini can now think using images as part of their reasoning process. Uploaded visuals can also be handled even if blurry or rotated — with the models able to adjust them using its own tools. https://t.co/1Z1waVkh7v” / X
https://x.com/rowancheung/status/1912561386208825751
OpenAI in talks to pay about $3 billion to acquire startup Windsurf https://www.cnbc.com/2025/04/16/openai-in-talks-to-pay-about-3-billion-to-acquire-startup-windsurf.html
OpenAI Said to Be In Talks to Buy Windsurf for About $3 Billion – Bloomberg https://www.bloomberg.com/news/articles/2025-04-16/openai-said-to-be-in-talks-to-buy-windsurf-for-about-3-billion?embedded-checkout=true
“ChatGPT has hit 1 billion weekly active users (WAU). OpenAI is not far from the vaunted 1 billion daily active users (DAU) club. https://x.com/bilawalsidhu/status/1911125917218508945
“All of your image creations, all in one place. Introducing the new library for your ChatGPT image creations—rolling out now to all Free, Plus, and Pro users on mobile and https://x.com/OpenAI/status/1912255254512722102
““Thinking with Images” has been one of our core bets in Perception since the earliest o-series launch. We quietly shipped o1 vision as a glimpse—and now o3 and o4-mini bring it to life with real polish. Huge shoutout to our amazing team members, especially: – @mckbrando, for” / X https://x.com/jhyuxm/status/1912562461624131982
“💥 o3 and o4-mini are launching today! Both models are mind-blowing. But maybe the coolest for me has been seeing them use tools as they think. They can search, write code, and manipulate images in the chain of thought, and it’s a huge multiplier. I will never forget the first” / X https://x.com/kevinweil/status/1912554045849411847
“Introducing OpenAI o3 and o4-mini—our smartest and most capable models to date. For the first time, our reasoning models can agentically use and combine every tool within ChatGPT, including web search, Python, image analysis, file interpretation, and image generation. https://x.com/OpenAI/status/1912560057100955661
“OpenAI o3 and o4-mini are our first models to integrate uploaded images directly into their chain of thought. That means they don’t just see an image—they think with it. https://x.com/OpenAI/status/1912560060284502016
“Just released o3 and o4-mini! These models feel incredibly smart. We’ve heard from top scientists that they produce useful novel ideas. Excited to see their positive impact on people’s daily lives and humanity’s hardest problems!” / X https://x.com/gdb/status/1912575762483540322
Introducing OpenAI o3 and o4-mini | OpenAI https://openai.com/index/introducing-o3-and-o4-mini/
“really good summary of o3’s strengths https://x.com/aidan_mclau/status/1912580976456474812
“One last note: we’ll also begin deprecating GPT-4.5 Preview in the API today as GPT-4.1 offers improved or similar performance on many key capabilities at lower latency and cost. GPT-4.5 in the API will be turned off in three months, on July 14, to allow time to transition (and” / X https://x.com/OpenAIDevs/status/1911860805810716929
OpenAI is building a social network | The Verge https://www.theverge.com/openai/648130/openai-social-network-x-competitor
Introducing GPT-4.1 in the API | OpenAI https://openai.com/index/gpt-4-1/
“gpt-4.1 is for developers:” / X https://x.com/OpenAIDevs/status/1912241877199581572
“GPT-4.1 outperforms GPT-4.5 in coding https://x.com/scaling01/status/1911828552452112536
“introducing the gpt-4.1 series. incredible at coding and instruction following, with a 1M token context window https://x.com/stevenheidel/status/1911830165317173740
“a lot of people were interested in how we made GPT-4.5 and what comes next. we did a podcast with alex paino, dan selsam, and @atootoon who helped drive the project. full episode coming soon, but here are some interesting clips:” / X https://x.com/sama/status/1910363426972635455
OpenAI’s Latest Breakthrough: AI That Comes Up With New Ideas — The Information https://www.theinformation.com/articles/openais-latest-breakthrough-ai-comes-new-ideas
Jerry Tworek on X: “Scaling is incredibly hard and demanding and leaves very little room for error in every little part of the training stack But once it works, it’s beautiful to see it https://t.co/I13hW5gAuE” / X
https://x.com/MillionInt/status/1912568397419954642
“the ability of the new models to effectively use tools together has somehow really surprised me intellectually i knew this was going to happen but it hits different to see it” / X https://x.com/sama/status/1912564175253172356
“We did not “solve math”. For example, our models are still not great at writing proofs. o3 and o4-mini are nowhere close to getting International Mathematics Olympiad gold medals.” / X https://x.com/polynoamial/status/1912575974782423164
“Amazon released a speech-to-speech AI called “Nova Sonic” Available on Bedrock, the model generates speech with a 1.09 sec latency, outperforming the best OpenAI models at 20% cost Amazon also launched Reel 1.1 AI for extended 2-min video generations https://x.com/adcock_brett/status/1911450262977368259
“Excited about the launch of Amazon Nova Sonic, our new speech-to-speech model that helps make AI voice applications feel remarkably natural. It’s designed to understand not just what people say, but how they say it – working with tone, style, and conversation flow including https://x.com/ajassy/status/1909691335877312757
“heard from some startup engineers that they lost several work hours gawking, stupefied, after they plugged 4.1 mini/nano into every previously-expensive part of their stack you can just do gpt-4o-quality things 25 × cheaper now” / X https://x.com/aidan_mclau/status/1911850291026362426
“o3 as a management coach, for personalized learning, and more: https://x.com/gdb/status/1912568418626420804
Our updated Preparedness Framework | OpenAI https://openai.com/index/updating-our-preparedness-framework/
“Alright, now that I manually built from scratch an AI Agent running on MCP with tool calling, I’ll scale the approach by using OpenAi’s new Agent SDK. If your are interested in building AI Agents, it definitly has benefits, while still being 100% free https://x.com/aranimontes/status/1907482776799949152
“I didn’t just build an outbound agent. I went a step further and — I cloned 11x. It finds leads, qualifies them, replies, and updates our Notion CRM — fully autonomous, 24/7. Costs less than your Slack plan. $0.01 per lead. Zero humans. Built using @OpenAI’s Agents SDK and https://x.com/n_sri_laasya/status/1905402229823324231
“🚀 Built a Simple AI Agent using UV, OpenAI Agents SDK & Gemini! 🎉 Big thanks to Sir @0xAsharib for the guidance! 🙌 #AI #OpenAI #Gemini #RamadanCodingNights #Tech #pythonlearning #agent #python #programminghelp https://x.com/tayyeba_ali/status/1905805581794861084
“After using them both, I think that Gemini 2.5 & o3 are in a similar sort of range (with the important caveat that more testing is needed for agentic capabilities) Each has its own quirks & you will likely prefer one to another, but there is a gap between them & other models” / X https://x.com/emollick/status/1912573923163504704
“I’m really digging 4o → @v0 for design → app. 4o can help you “design outside the box”. Prompting is the future. https://x.com/rauchg/status/1911141908187058219
“”at or near genius level”” / X https://x.com/sama/status/1912558996084650003
“a few times a year i wake up early and can’t fall back asleep because we are launching a new feature ive been so excited about for so long. today is one of those days!” / X https://x.com/sama/status/1910334443690340845
“codex cli: https://x.com/sama/status/1912586034568945828
“how about we fix our model naming by this summer and everyone gets a few more months to make fun of us (which we very much deserve) until then?” / X https://x.com/sama/status/1911906570835022319
“super appreciate everyone who spent their evening with us yesterday; the feedback was very helpful. i think we will be able to deliver something great!” / X https://x.com/sama/status/1910867196122931439
“we expect to release o3-pro to the pro tier in a few weeks” / X https://x.com/sama/status/1912558745013612888
“we have trained more models and they are good in some things” / X https://x.com/sama/status/1910800981429792878
“we’ve got a lot of good stuff for you this coming week! kicking it off tomorrow.” / X https://x.com/sama/status/1911490401221120284
Rowan Cheung on X: “Benchmarks and costs for o3 and o4-mini below. OpenAI said o3 hits SOTA performance across benchmarks for coding, real-world software tasks, and multimodality (MMMU). Some major efficiency leaps for both models compared to performance. https://t.co/BrmbeD0Z9N” / X
https://x.com/rowancheung/status/1912561391505977681
“o3 is out and it is absolutely amazing!! i’ve been playing with it for a week or so and it’s already my go-to model. it’s fast, agentic, extremely smart, and has great vibes. some of my top use cases: – it flagged every single time I sidestepped conflict in my meeting https://x.com/danshipper/status/1912551847056785841
“🚨New Tutorial Alert! Build an AI Agent to help solve the crypto UX problem—powered entirely by open source: – @OpenAI Agent SDK – @Lilypad_Tech and @ollama for AI Inference – web3py and eth-account for on-chain actions – @coingecko for crypto price feeds Blog link below👇 https://x.com/narb_s/status/1906738541583048868
“BREAKING: OpenAI just released o3 and o4-mini, the company’s two newest reasoning models. o3 is OpenAI’s ‘most advanced’ reasoner yet, while o4-mini brings strong results at lower costs. Also launching is Codex CLI, a new open-source coding agent. More info in thread below: https://x.com/rowancheung/status/1912561382832369824
“Meet Codex CLI—an open-source local coding agent that turns natural language into working code. Tell Codex CLI what to build, fix, or explain, then watch it bring your ideas to life. https://x.com/OpenAIDevs/status/1912556874211422572
“Also released today is Codex CLI — an open-source lightweight coding agent that runs in your terminal: https://x.com/gdb/status/1912576201505505284
“Just built an Email Sending Agent that works with natural language! You type → It drafts & sends the email for you. Powered by: → @OpenAI Agents SDK → @nebiusaistudio AI LLMs → @resend for email delivery Full step-by-step breakdown in my latest video! 👇 https://x.com/Arindam_1729/status/1906729998913786181
“o3 is amazing and obviously the best model, but underdelivers/sucks in some areas and is ridiculously marketed as “AGI” or “at or near genius level” or “fast, agentic”, when Gemini is faster and Sonnet is more agentic here are the things that suck for o3: – too expensive (4-5x” / X https://x.com/scaling01/status/1912633356895814019
Rowan Cheung on X: “OpenAI president Greg Brockman opened the stream by proclaiming today’s releases as a GPT-4 level “qualitative step into the future”. He also said top scientists have said the models produce “legitimately good and useful novel ideas”. Another leap forward today 📶” / X
https://x.com/rowancheung/status/1912561397361226043
“🚀Advance Agent Built under Sir @0xAsharib mentorship during #20_Days_Of_Ramadan_Coding 🌙 🔥 Tech Stack → UV + OpenAI Agents SDK + Chainlit 💡 Dual AI Power → OpenAI + Gemini 🔒 GitHub Auth | 📂 Persistent Chats | 💬 Smart UI#AI #RamadanCoding #MentorshipMatters #BuildInPublic https://x.com/MMaaz28220/status/1905587231420449066
“Coding with models like o4-mini and Gemini 2.5 Pro is a magical experience. Planning improved, code understanding is next level, and long code generation is more accurate. If you use agentic IDEs like Windsurf, reasoning models feel 100x smarter and more useful. Golden era!” / X https://x.com/omarsar0/status/1912878408280727632
“WebToAgent: Convert Entire Websites into Agents WebToAgent turns any website into a conversational assistant you can simply chat with to access information. WebToAgent is built on the top of Firecrawl and OpenAI Agent SDK. Key Features – Website content extraction: Crawl https://x.com/kalyan_kpl/status/1907771788060147900
“>be you >work in HFT shaving nanoseconds off latency or extracting bps from models >have existential dread >see this tweet, wonder if your skills could be better used making AGI >apply to attend this party, meet the openai team >build AGI” / X https://x.com/sama/status/1911910628232691947
“There seems to be a real bifurcation growing between the IT-facing API side of OpenAI & the user-facing ChatGPT side. You can increasingly only access the high power, high cost models through ChatGPT while the API supports cheap fast models. Some tension there for any AGI future” / X https://x.com/emollick/status/1911869865607995438
o3 and AGI, is April 16th AGI day? – Marginal REVOLUTION https://marginalrevolution.com/marginalrevolution/2025/04/o3-and-agi-is-april-16th-agi-day.html
“Interesting argument from @tylercowen. Is o3 good enough to be AGI? The counter argument might force us to wait until ASI, because only then will an AI definitively outperform all humans at all tasks. In the meantime we have a Jagged AGI, with subhuman & superhuman abilities. https://x.com/emollick/status/1912571045065736459
“I’ve seen enough, it’s AGI https://x.com/fabianstelzer/status/1912749181858357546
“o3 vs Sonnet 3.7 and Gemini 2.5 Pro GPQA: o3 – 83.3% Sonnet – 84.8% Gemini – 84.0% SWE-bench Verified: o3 – 69.1% (regression over preview score from December) Sonnet – 70.3% Gemini 63.8 % AIME 2024: o3 – 91.6 % Sonnet – 80.0% Gemini – 92.0% Aider: o3 – 81.3 % Sonnet – 64.9 %” / X https://x.com/scaling01/status/1912568851604119848
“GPT-4.1 benchmark results: GPT-4.1 scores worse than GPT-4, Opus and Llama-3.1-70B (lol) GPT-4.1 API version is WORSE than Optimus Alpha and Quasar Alpha (so results of quasar were just fluff hype) GPT-4.1 mini scores worse than Qwen2.5 32B, Llama-4 Maverick and Claude 3 Haiku https://x.com/scaling01/status/1911847193465471374
“it’s finally out! OAI has an answer to Claude Code – and this time it’s fully Apache 2 open source!! https://x.com/swyx/status/1912558096553242663
“one of my favorite benchmarks: WeirdML GPT-4.1-mini outperforms GPT-4.1 it ranks 6th in the overall ahead of Grok-3, DeepSeek-R1 and Sonnet 3.5 I have seen GPT-4.1-mini overperforming relative to GPT-4.1 in other benchmarks too. GPT-4.1-mini might be the hidden gem of this” / X https://x.com/scaling01/status/1912117156751229268
“Is there anything more frustrating that dictating a multi-minute voice note and the chatgpt app retorts with a “sorry, i didn’t quite catch that”” / X https://x.com/bilawalsidhu/status/1912189883545686202
“New model in our API — GPT-4.1. It’s great at coding, long context (1 million tokens), and instruction following.” / X https://x.com/gdb/status/1911837358263292409
“GPT-4.1 (and -mini and -nano) are now available in the API! these models are great at coding, instruction following, and long context (1 million tokens). benchmarks are strong, but we focused on real-world utility, and developers seem very happy. GPT-4.1 family is API-only.” / X https://x.com/sama/status/1911830886896799931
“I had early access, o3 is an impressive model, seems very capable. Some fun examples: 1) Cracked a business case I use in my class 2) Creating some SVGs (images created by code alone) 3) Writing a constrained story of two interlocking gyres 4) Hard science fiction space battle. https://x.com/emollick/status/1912552106214502739
“ChatGPT Plus, Pro, and Team users will see o3, o4-mini, and o4-mini-high in the model selector starting today, replacing o1, o3-mini, and o3-mini-high. ChatGPT Enterprise and Edu users will gain access in one week. Rate limits across all plans remain unchanged from the prior set” / X https://x.com/OpenAI/status/1912560062004179424
“OpenAI released GPT-4.1, 4.1 Mini, and the ultra-fast 4.1 Nano—all designed for devs — Each model beats GPT-4o and 4o mini on dev tasks — 1M token context windows — GPT-4.1 scored 55% on SWE-Bench Verified — Starting at $0.10/0.40 per million I/O tokens https://x.com/rowancheung/status/1912034453771202847
“BOOM! Psyched to welcome @cohere on the Hub -providing blazingly fast inference for their Open and Enterprise models! 🔥 Starting today you can access all open Cohere models directly from the Hub with OpenAI compatibility too! Bonus: The models are really really good at Tool https://x.com/reach_vb/status/1912523838723662064
“Our new @OpenAI o3 and o4-mini models further confirm that scaling inference improves intelligence, and that scaling RL shifts up the whole compute vs. intelligence curve. There is still a lot of room to scale both of these further. https://x.com/polynoamial/status/1912564068168450396
“Something magic about these charts with compute on the x-axis, and clear steady improvements (across many metrics) on the y-axis. Really makes you feel like the field is uncovering some underlying fundamental laws of intelligence.” / X https://x.com/gdb/status/1912590623955382445
“@OpenAI Maybe one reason GPT-4.1 is not available on ChatGPT is to incentivize college students to subscribe. Since the free GPT-4.1 mini matches the paid GPT-4.1 too closely for key users (e.g., students doing homework), it could remove the incentive to subscribe to ChatGPT Plus.” / X https://x.com/DanHendrycks/status/1911837235521163670
“Yesterday was crazy so I missed my post but damn the 4.1s awesome. @michpokrass @jhyuxm @johnohallman @SuvanshSanjeev @StrongDuality and so many others did a phenomenal job. We’re truly horrible at naming but the secret trick is that the models with mini in their name are 🔥” / X https://x.com/_aidan_clark_/status/1912191545203413419
“The hottest new programming language is Lean ⚡️ Kimina-Prover beats Gemini 2.5 Pro and o3-mini on Olympiad-level math with just 7B parameters! The application of RL with tool-use (Lean compiler) looks like a blueprint of what’s coming with code agents this year 🔥 https://x.com/_lewtun/status/1911793153931100180
“was in rome recently and, in one CoT, o3: >reasoned hard >resized the image and zoomed in >searched the internet several times >figured out my location >checked memory; deduced i was on vacation gave me a better explanation than the museum lol actually blew my mind https://x.com/aidan_mclau/status/1912560625005522975
“We launched o3 and o4-mini today! Reasoning models are so much more powerful once they learn how to use tools end-to-end. Some of the biggest lifts are coming in multimodal domains like visual perception (see how it solves a maze in our blogpost: https://x.com/markchen90/status/1912609299270103058
“Announcing GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano in the API. TL;DR: Major improvements on coding, instruction following, and long context. 💥 00:00 Intro 02:18 Coding 04:53 Instruction following 06:58 Long context 10:22 Demos, pricing, and availability 20:00 @windsurf_ai https://x.com/OpenAIDevs/status/1911859780236214426
“Resisting the standard urge to tweet about OpenAI’s naming system given today’s new products are named 4.1, 4.1-mini, and 4.1-nano, given the existence of 4o, the upcoming o4, and the existing 4.5. Don’t worry, the numbering system is completely uninformative as to capabilities.” / X https://x.com/emollick/status/1911861186200469932
“OpenAI’s GPT-4.1 series is a solid upgrade: smarter and cheaper across the board than the GPT-4o series @OpenAI’s GPT-4.1 family includes three models: GPT-4.1, GPT-4.1-mini and GPT-4.1 nano. We have independently benchmarked these with our Artificial Analysis Intelligence Index https://x.com/ArtificialAnlys/status/1912177623360479281
OpenAI o3 & o4-mini – YouTube https://www.youtube.com/watch?v=sq8GBPUb3rk
“THIS IS HUGE o4-mini is CHEAPER AND BETTER across the board https://x.com/scaling01/status/1912553316849942626
“💥 New today in the API: GPT 4.1, GPT 4.1 mini, and GPT 4.1 nano. These models are great at coding (54 on SWE-bench verified for a non-reasoning model!) and instruction following, and all three offer 1M context. We’ve also dropped prices: GPT 4.1 is 26% cheaper than 4o, and 4.1” / X https://x.com/kevinweil/status/1911833354682401148
“o3 and o4-mini are out! they are very capable. o4-mini is a ridiculously good deal for the price. they can use and combine every tool within chatgpt. multimodal understanding is particularly impressive.” / X https://x.com/sama/status/1912558064739459315
“🚨 @OpenAI has launched o3 and o4-mini! 🎉 o3 is absolutely dominating the SEAL leaderboard with #1 rankings in: 🥇: HLE 🥇: Multichallenge (multi-turn) 🥇: MASK (honesty under pressure) 🥇: ENIGMA (puzzle solving) Congrats @sama @markchen90 & team 🔗: https://x.com/alexandr_wang/status/1912555697193275511
Details about METR’s preliminary evaluation of OpenAI’s o3 and o4-mini | METR’s Autonomy Evaluation Resources https://metr.github.io/autonomy-evals-guide/openai-o3-report/
“What burning question shall we ask a supersmart AI? “o3, i need a chart that compares who would win in a fight among a variety of historical figures, cheeses, and small animals, each cell should represent a match up between x and y combatants, don’t repeat yourself.” https://x.com/emollick/status/1912607439985221881
“GPT-4o and GPT-4o mini will remain available in the API. While we plan to continue supporting them, we think most developers will benefit from switching to GPT-4.1. Here’s a prompting guide to help with the transition: https://x.com/OpenAIDevs/status/1911860803944271969
“OpenAI have announced the availability of GPT-4.1 in the API, and of course we have day 0 support! To get it just install the latest version of our OpenAI integration: pip install -U llama-index-llms-openai Learn more here: https://x.com/llama_index/status/1911863053257445713
“Time for another vibe check” / X https://x.com/bilawalsidhu/status/1912600770043609476
“O3 is a very strong model, but still a jagged one.” / X https://x.com/emollick/status/1912565658233061385
“A big shoutout to the tireless babysitters who saw o3 to fruition – this includes @zhouwenda and @mckbrando who you saw on stream, but also @alexwei_, @bminaiev, @mikegmalek, @ilge, Botao, Vineet, Hunter, and many many others behind the scenes.” / X https://x.com/markchen90/status/1912616143061414222
“Thanks to everyone who joined our first Developer Listening Session in SF! It was inspiring to see the community come together to help shape our approach to the next open model. Your feedback is invaluable—more to come soon! https://x.com/OpenAIDevs/status/1910863030537138275
“We tested a pre-release version of o3 and found that it frequently fabricates actions it never took, and then elaborately justifies these actions when confronted. We were surprised, so we dug deeper 🔎🧵(1/) https://x.com/TransluceAI/status/1912552046269771985
“These models can navigate large codebases and generate novel ideas. Tool use makes these models a lot more useful. The o-series of models is now combined with their full suite of tools.” / X https://x.com/omarsar0/status/1912554367711957437
“sama has executed a 200iq joke here that all the replies failed because they only llm everyone in HFT knows exactly what i means to advertise availability of something that is suddenly not there when you want it top tier hiring move” / X https://x.com/swyx/status/1911918989464461663
“And GPT-4.1 makes four (I am not 100% sure this came out fully as the AI conceived, but it is only the fourth model to create a shader that runs in twigl the first time around, and it is probably the most visually complex) https://x.com/emollick/status/1911967773489512466
“This is true. It was big news in June 2023 when a paper said that GPT-4 could get between 90%-100% on the problems in the MIT core CS & math classes. It later turned out to be wrong I think Ben is right that current models could plausibly do this. But it is no longer a big deal.” / X https://x.com/emollick/status/1910201610875043991
“Our latest @OpenAI model, GPT-4.1, achieves 55% on SWE-Bench Verified *without being a reasoning model*. @michpokrass and team did an amazing job on this! (New reasoning models coming soon too.) https://x.com/polynoamial/status/1911831926241153170
“GPT-4.1 is an interesting model to me. I believe the next challenge with AI isn’t about training bigger or better models, but about figuring out what problems we really want AI to solve, and how we measure its real-world impact. For years, progress was all about coming up with” / X https://x.com/skirano/status/1912156805901205986
“o3 seems to hallucinate >2x more than o1, according to the system card so hallucinations could scale *inversely* with increased reasoning (unlike for increased model size), bc outcome-based optimization incentivizes confident guessing (the Transluce example is kinda hilarious) https://x.com/ryan_t_lowe/status/1912641520039260665
API Organization Verification | OpenAI Help Center https://help.openai.com/en/articles/10910291-api-organization-verification
“OK, pretty impressive showing from GPT-4.1 in the p5js space ship control panel generation challenge. Big gains from GPT-4 (scroll down) and impressive overall (I did add “make it elaborate” to the prompt as some models are lazier than others, otherwise the prompt is unchanged) https://x.com/emollick/status/1911966088339894669
ChatGPT became the most downloaded app globally in March | TechCrunch https://techcrunch.com/2025/04/11/chatgpt-became-the-most-downloaded-app-globally-in-march/
“According to @windsurf_ai, GPT-4.1 performed really well (a 60% improvement over GPT-4o) on internal benchmarks like the SWE-benchmark that validates end-to-end performance. – GPT-4.1 reduces # of times it needs to read unnecessary files by 40% compared to other models. – It” / X https://x.com/omarsar0/status/1911870478857437540
“@Replit @v0 2) We built an ETL script in Replit to extract key features from each release using the OpenAI API https://x.com/neondatabase/status/1909047458271015007
“After spending the past ~1 year building conviction in what I want to do next, I’ve joined the research team @OpenAI! I’ve been here for 2 weeks, and am so energized / convinced there’s no better place to be in the world right now. I’m working as a resident under @BorisMPower https://x.com/MajmudarAdam/status/1911821179960393963
“the openai team is executing just ridiculously well at so many things right now, the coming months and years should be amazing (a lotta stuff is messy and very broken too of course)” / X https://x.com/sama/status/1912300398263824749
“@windsurf_ai GPT-4.1 also follows instructions more reliably. On internal evals, GPT-4.1 outranks GPT-4o on tasks like format adherence, complying with negative instructions, and ordering. https://x.com/OpenAIDevs/status/1911860099829674184
“openai usage has gone nuts over the past month. but a surprisingly cool thing is you can ask the internet “hey can we please get hundreds of thousands of more GPUs quickly” and…the internet will deliver. 🩵” / X https://x.com/sama/status/1910398997333970984
“if you’re confused about the GPT naming scheme, realize that it’s actually GPT-4.10 so obviously it comes after GPT-4.5, it makes perfect sense!” / X https://x.com/iScienceLuvr/status/1911832534796886439
“do not use GPT-4.1-nano it is a terrible model” / X https://x.com/scaling01/status/1911852197714731276
“we have greatly improved memory in chatgpt–it can now reference all your past conversations! this is a surprisingly great feature imo, and it points at something we are excited about: ai systems that get to know you over your life, and become extremely useful and personalized.” / X https://x.com/sama/status/1910380643772665873
“@sama openai hunting cap is a must for the next podcast https://x.com/willdepue/status/1911591028697833779
“people loved the GPT-4.5 podcast! what would you like to see one on next?” / X https://x.com/sama/status/1910861296217522610
“I’m absolutely blown away by @OpenAI’s new o3 model! I’ve had early access and haven’t put it down for days. This release feels like the milestone we experienced with o1-preview and o1-pro, but smarter and more reliable in every way, it truly cranks everything up to eleven! In https://x.com/DeryaTR_/status/1912558350794961168
“GPT-4.1 in the API https://x.com/OpenAI/status/1911824315194192187
“@MeansTestRule @OpenAI @michpokrass We are really bad at naming things and I think at this point we’ve given up trying” / X https://x.com/polynoamial/status/1911843302770643004
“gpt-4.1-nano is the cheapest and fastest model we’ve ever released. it’s just $0.10/1M input ($0.03 cached) and $0.40/1M output https://x.com/stevenheidel/status/1911830168118923291
“OpenAI is getting very good at creating viral moments.” / X https://x.com/emollick/status/1910771755771167121
“Friends and family often ask whether LLMs will replace search engines. I don’t think so because LLM-based search wouldn’t work without search engines. General audience LLMs like GPT-4o literally rely on them to answer many queries. In that sense, LLMs are (for now, a very” / X https://x.com/rasbt/status/1911467070975271217
“The BLS just released new OEWS data, so we can actually update this (it’s an annual survey every May — so this is 18 months of data post-ChatGPT, instead of 6 months) From 2023 -> 2024: (1) Translator real wages +0.8% (2) Translator employment/employment share slightly up ⬆️ https://x.com/BasilHalperin/status/1912268400530739254
“”o3, make me a movie i can download that involves an otter and an airplane. figure out how to do it with the tools you have.” o3 has no movie capability, so It improvises decides to draw each frame and then stitch them together into a GIF to download, this was all first shot https://x.com/emollick/status/1912597487287705965
“many people worked on this, but @michpokrass really drove it. she is truly amazing, and one of the rare people who can do everything from deeply understand what users want to complex details of research and everything in between. although rare, these people make magic happen!” / X https://x.com/sama/status/1911831955441582545
“It often takes hours for me to track down bugs in my code that result in subtly incorrect numbers in the output. I’ve started asking o3 to just “look for bugs in this” and it catches like 80% of bugs before I run anything. It’s also great at one-shotting standalone scripts and” / X https://x.com/itsclivetime/status/1912569732693438771
“@windsurf_ai The GPT-4.1 family is a major upgrade over GPT-4o for real-world software engineering work. It’s significantly more skilled at frontend coding, suggesting clean diffs, structured responses, reliable tool use, and more. https://x.com/OpenAIDevs/status/1911859923161428002
“Sam Altman on GPT-5 in Berlin today “How many of you think you will still be smarter than GPT-5 ? I don’t think I’m gonna be smarter than GPT-5” #Gpt5 @sama 👀 https://x.com/RaphaelDabadie/status/1887833328096579625
“Social media captures what we post and consume. ChatGPT captures what we think and feel. As Sam Altman said at TED yesterday — the upload happens bit by bit. Layer in AR glasses and always-on ambient assistants, and these AI systems will know us better than we know ourselves.” / X https://x.com/bilawalsidhu/status/1911196138281001177
OpenAI looked at Cursor before considering deal with rival Windsurf https://www.cnbc.com/2025/04/17/openai-looked-at-cursor-before-considering-deal-with-rival-windsurf.html
“Harvard study shows we can measure leadership skills by seeing how folks manage GPT-4o simulated people AI assessments strongly correlate (r=0.81) with human team assessments. Effective leaders ask questions & do conversational turn-taking and have fluid & social intelligence https://x.com/emollick/status/1910321275848831336
“Large scale job displacements due to AI are likely to occur more slowly than a lot of people talking about AI and work might suspect. Translators were paid more and in more demand a year after GPT-4 than they were the year it launched.” / X https://x.com/emollick/status/1912272981692240104
“GPT-4.1 still underperforming DeepSeekV3 in Coding but 8x more expensive” / X https://x.com/scaling01/status/1911830809679368248
“GPT-4.1 underperforming DeepSeek-V3-0324 by over 10% on AIME (also slightly underperforming on GPQA) at 8x the price” / X https://x.com/scaling01/status/1911831700964872531
“GPT-4.1 Prompting Guide from OpenAI’s official Github Repo 1. Clarity and Context Always provide clear, specific instructions. Short, direct prompts can work, but adding examples and explicit goals greatly enhances performance. If the model’s behavior seems off, add a single https://x.com/rohanpaul_ai/status/1911940922448945488
12 former OpenAI employees asked to be heard in Elon Musk’s lawsuit against the company; one calls Sam Altman a ‘person of low integrity’ | Fortune https://fortune.com/2025/04/11/12-ex-openai-employees-elon-musk-lawsuit-amicus-filing-lessig/
Group of ex-OpenAI employees back Musk’s lawsuit to halt OpenAI restructure | Reuters https://www.reuters.com/technology/artificial-intelligence/group-ex-openai-employees-back-musks-lawsuit-halt-openai-restructure-2025-04-12/
“Sam Altman confirms: OpenAI is planning to release a very powerful open-source model. It could be “near the frontier” and better than any current open-source model out there. https://x.com/slow_developer/status/1911092611751850330
“Its hard to understand how ChatGPT’s extended memory works. It is pretty good at playing at cold readings that sound profound, but seems to struggle with useful details & patterns from chats. Was there any info released about whether it is RAG, some sort of summary file, etc?” / X https://x.com/emollick/status/1910890927595418015
“THIS IS BAD NEWS o3 is worse at replicating research papers than o1 https://x.com/scaling01/status/1912554822454116736
“cute little robot” / X https://x.com/sama/status/1910843661123809627
“ignore literally all the benchmarks the biggest o3 feature is tool use ofc it’s smart, but it’s also just way more useful >deep research quality in 30 seconds >debugs by googling docs and checking stackoverflow >writes whole python scripts in its CoT for fermi estimates” / X https://x.com/aidan_mclau/status/1912559163152253143
