➀ Microsoft has unveiled Phi-4, a small AI model with impressive mathematical reasoning capabilities; ➁ Phi-4 outperforms larger models and even beats heavyweights like Claude Sonnet 3.5, GPT 4o, and Google Gemini Pro 1.5 in math competition problems; ➂ Microsoft's advancements in developing Phi-4, including the use of high-quality synthetic data sets and innovative post-training methods, push the frontier of size vs quality.
Related Articles
- Not to be outdone by ChatGPT, Microsoft Copilot humiliates itself in Atari 2600 chess showdown — another AI humbled by 1970s tech despite trash talk2 months ago
- AI taught to analyze Windows crash dumps, released to open source — 'like going from hunting with a stone spear to using a guided missile'4 months ago
- Nuclear Power: The Future To Satisfy Energy-Hungry Data Centres5 months ago
- AI drives VC funding5 months ago
- The Big Dilemma5 months ago
- Snapdragon X exclusive Copilot+ features begin trickling through to modern x86 Windows 11 PCs5 months ago
- Microsoft makes ChatGPT's most-advanced $200 a month AI free7 months ago
- Trump announces 0.5trn investment in AI8 months ago
- Google CEO fires one off at Microsoft by saying its ready for the AI race 'any day, any time'9 months ago
- Microsoft Pinky-Promises It's Not Training AI On Your Word And Excel Data9 months ago