Grok has the best value for coding
Grok has the best value for codingCognition: The FrontierCode leaderboard is now live: a dedicated page that tracks which models are writing code you’d actually merge.…
X / Twitter 6832 stories
in line with my impression so far 5.5, not 5.6 But holy shit, 5.5 is not that old
in line with my impression so far5.5, not 5.6But holy shit, 5.5 is not that oldArtificial Analysis: Kimi K3 in Kimi Code CLI scores 57 and…
RT cheaty: what am i even looking at? is this a joke? how was this released two months ago and you're only posting abotu it now? the weights have been…
RT cheatywhat am i even looking at? is this a joke? how was this released two months ago and you're only posting abotu it now? the…
Underrated take Yes, "naturally" Chinese labs should stick to models on the level of DSV4-Flash. Super cheap, decent-ish. Economy class. Spend the sur…
Underrated takeYes, "naturally" Chinese labs should stick to models on the level of DSV4-Flash. Super cheap, decent-ish. Economy class. Spend the surplus on marketing and "Application…
The funny thing is if K3 is heavily distilled from *Opus* it dunks on
The funny thing is if K3 is heavily distilled from *Opus* it dunks onAlex Kolicich: It looks likely that Kimi K3 was heavily distilled from FableIf…
btw if you havent set your {codex | claude | gemini | devin} automations to autoresearch how to improve your seo/aeo every week you are really truly m…
btw if you havent set your {codex | claude | gemini | devin} automations to autoresearch how to improve your seo/aeo every week you are really…
this is cool: https://welcome-to-codex.openai.chatgpt.site/
this is cool:https://welcome-to-codex.openai.chatgpt.site/
Re Here's how to add the Codex Security plugin in Codex and get started: Add the plugin in Codex. After installation is complete, the button changes t…
Re Here's how to add the Codex Security plugin in Codex and get started:Add the plugin in Codex. After installation is complete, the button changes to…
GPT-5.6 Sol sets a new state of the art in cybersecurity on “The Last Ones” cyber range. We’re already seeing that capability translate into defens…
GPT-5.6 Sol sets a new state of the art in cybersecurity on “The Last Ones” cyber range.We’re already seeing that capability translate into defensive outcomes: helping…
RT Min Choi: Grok 4.5 is only 1.5T parameters. Kimi K3 is 2.8T… and costs 3× more per task. SpaceXAI cracked intelligence efficiency. Now imagine G…
RT Min ChoiGrok 4.5 is only 1.5T parameters.Kimi K3 is 2.8T... and costs 3× more per task.SpaceXAI cracked intelligence efficiency.Now imagine Grok at 3T.Elon Musk: Grok…
it's funny how @TMTLongShort's bio starts with "Economic Blitzkrieg". Unfortunate naming. I wonder if Nazis also thought they're losing because they'r…
it's funny how @TMTLongShort's bio starts with "Economic Blitzkrieg". Unfortunate naming. I wonder if Nazis also thought they're losing because they're "pussyfooting around", being too Merciful…
RT DeepBurner: Added kimi-3 to my Surface Evolver Bench. It scores on par with Fable 5 and Sol (xhigh). The cost is in fact higher than Sol, as kimi i…
RT DeepBurnerAdded kimi-3 to my Surface Evolver Bench. It scores on par with Fable 5 and Sol (xhigh). The cost is in fact higher than Sol,…
RT Laude Institute: Andy @andykonwinski and @swyx talked through why AI researchers may be uniquely well-positioned to wield power responsibly right n…
RT Laude InstituteAndy @andykonwinski and @swyx talked through why AI researchers may be uniquely well-positioned to wield power responsibly right now, why most of them are…
RT Rothmus 🏴: On this day in 1918 the Bolsheviks herded Tsar Nicholas II, Empress Alexandra, and their five children into a basement in Yekaterinbu…
RT Rothmus 🏴On this day in 1918 the Bolsheviks herded Tsar Nicholas II, Empress Alexandra, and their five children into a basement in Yekaterinburg and opened…
The bagholder mindset is fearsome to behold > We have already boarded a train we cannot get off—one that leaves us no choice but to keep pouring mone…
The bagholder mindset is fearsome to behold> We have already boarded a train we cannot get off—one that leaves us no choice but to keep pouring…
RT AfterQuery: Kimi K3 ranks #1 on @AfterQuery's SpreadsheetBench 2, surpassing Claude Fable 5. An open weight model now outperforms all closed-source…
RT AfterQueryKimi K3 ranks #1 on @AfterQuery's SpreadsheetBench 2, surpassing Claude Fable 5. An open weight model now outperforms all closed-sourced models. Read more in the…
Sol gets the thing done:
Sol gets the thing done:tobi lutke: OpenAI shipped /goal not too long ago. I feel like GPT-5.6-sol is the first model that just doesn't need it…
Still building your #SIGGRAPH2026 agenda? We've got you covered. 🤝 👉 Learn the latest research breakthroughs at the NVIDIA keynote. 👉 Build n…
Still building your #SIGGRAPH2026 agenda? We've got you covered. 🤝👉 Learn the latest research breakthroughs at the NVIDIA keynote.👉 Build new skills in hands-on labs.👉 Explore…
China is really obsessed with material science they graduate buttloads of these researchers, publish like 60% of the world's total high impact materia…
China is really obsessed with material sciencethey graduate buttloads of these researchers, publish like 60% of the world's total high impact material science and engineering researchand…
GPT-5.6 Sol is the state of the art in cyber. Seeing significant results in applying it to finding and fixing novel vulnerabilities. Sign up as a defe…
GPT-5.6 Sol is the state of the art in cyber. Seeing significant results in applying it to finding and fixing novel vulnerabilities.Sign up as a defender…
The issue isn't so much about logits, this is a red herring imo the main problem is on-policy vs off-policy RL. Chinese labs are shifting towards MOPD…
The issue isn't so much about logits, this is a red herring imothe main problem is on-policy vs off-policy RL. Chinese labs are shifting towards MOPD.…
10,000 reasons people love GPT-5.6 Sol https://switch-to-codex.openai.chatgpt.site/
10,000 reasons people love GPT-5.6 Sol https://switch-to-codex.openai.chatgpt.site/Tibo: Or… what if we gave you $100 in Codex credits if you tell us what you love about GPT-5.6…
I really want to see more info on this new Ascend configuration @zephyr_z9 do you hear anything from WAIC?
I really want to see more info on this new Ascend configuration@zephyr_z9 do you hear anything from WAIC?Kyle Chan: China’s message at this year’s World AI…
it's remarkable how people don't reflect on the way differential credulity shapes their views. Some shitty handwavy Reuters report: a credible source …
it's remarkable how people don't reflect on the way differential credulity shapes their views.Some shitty handwavy Reuters report: a credible source on "real" Chinese position on…