LiveCodeBench Pro @SentientAGI Serious Reality Check for AI Coders 🔷I just took a look at LiveCodeBench Pro, a new benchmark developed by Sentient and unveiled at NeurIPS and honestly, this feels like the first truly serious test of AI’s real programming ability. 🔷Unlike older benchmarks that were often too lenient (many leaked tasks or could be gamed with clever prompting). LCB-Pro from Sentient follows the rigorous standards of Codeforces / ICPC / IOI featuring real competitive problems, strict time and memory limits, and even a “hack phase” to catch hidden bugs.

Google DeepMind’s Gemini 2.5 Achieves Gold Medal in International Programming Competition : A specialized version of Google DeepMind’s Gemini 2.5 AI model became the first AI to win a gold medal at the International Olympiad in Informatics (IOI), solving complex real-world programming problems that challenged human competitors. This marks a “historic” breakthrough in AI problem-solving capabilities. Source: The Guardian (Published September 17, 2025).

OpenAI’s AI Surpasses Human and Google Competitors at ICPC World Finals An AI system by OpenAI achieved a historic milestone at the 2025 International Collegiate Programming Contest (ICPC) World Finals, solving all 12 tasks and outperforming both human teams and Google DeepMind’s Gemini 2.5 Deep Think model. Competing under standard contest conditions, OpenAI’s ensemble—featuring GPT-5 and an experimental reasoning model—submitted solutions within the five-hour time limit. GPT-5 provided correct answers for 11 problems; the experimental model solved the most challenging twelfth problem after multiple attempts. In comparison, Gemini 2.5 Deep Think completed 10 problems, missing two tasks. OpenAI highlights this victory as proof of their models’ advanced reasoning capabilities and general problem-solving excellence, echoing previous gold-level results at the IMO and IOI.

Grok 4 dominates IOI benchmark with 26.2% accuracy! Is this the start of a new AI champion? • GPT-5: 20.0% • Gemini 2.5 Pro: 17.1% Who will win the AI race? #AI #Grok4 #GPT5 #Gemini

Grok 4 leads the IOI benchmark! Who will win the AI accuracy race? • Grok 4: 26.2% • GPT-5: 20.0% • Gemini 2.5 Pro: 17.1% #AI #Benchmark #Grok4 #GPT5

Was kinda disappointed when sama mentioned gpt-5 will essentially be a routing model. Was at least expecting it to also be another hero training run Doesn’t feel like a routing model deserves the gpt-5 labelI am excited for the model that’s winning IMO/IOI though. They explicitly said it’s not GPT-5

Was kinda disappointed when sama mentioned gpt-5 will essentially be a routing model. Was at least expecting it to also be another hero training run Doesn’t feel like a routing model deserves the gpt-5 labelI am excited for the model that’s winning IMO/IOI though. They explicitly said it’s not GPT-5

Was kinda disappointed when sama mentioned gpt-5 will essentially be a routing model. Was at least expecting it to also be another hero training run Doesn’t feel like a routing model deserves the gpt-5 labelI am excited for the model that’s winning IMO/IOI though. They explicitly said it’s not GPT-5

Was kinda disappointed when sama mentioned gpt-5 will essentially be a routing model. Was at least expecting it to also be another hero training run Doesn’t feel like a routing model deserves the gpt-5 labelI am excited for the model that’s winning IMO/IOI though. They explicitly said it’s not GPT-5

Was kinda disappointed when sama mentioned gpt-5 will essentially be a routing model. Was at least expecting it to also be another hero training run Doesn’t feel like a routing model deserves the gpt-5 label I am excited for the model that’s winning IMO/IOI though. They explicitly said it’s not GPT-5

价格行为-信号k如入场k 多k线反转,iii形态,ioi形态 https://youtu.be/veaGcrpO9gY?si=Gy67AR04FxtcbSv- 来自 @YouTube b站:【价格行为-信号k如入场k 多k线反转,iii形态,ioi形态-哔哩哔哩】 https://b23.tv/xgLLNjV

ioi

Stealth consumer AI startup with IOI gold medalists, hiring: — Full-stack SWE — ML Engineer — ML Researcher Built something? DM your top build, hardest problem cracked, and links (GitHub / X / arXiv) Not you? $10k referral bonus if your intro joins

Hi, it’s 2045 and I am Parzival, a simple guy on a mission to find Dan Romer Halliday clues to save Farcaster before the evil IOI(X) corp come and take control of the last place where u can truly feel urself online. Today I woke up and logged in the latest FarcasterOS with my armored Farcaster Haptic suit. In the morning I usually do some virtual work on my Farm on /farville and then I take a rest sippin matcha w my /gmatcha frens. …