Microsoft's most capable new Phi 4 AI model rivals the performance of far larger systems
Microsoft, the American multinational technology corporation, has launched a new artificial intelligence (AI) model named Phi 4. This latest Microsoft launch includes several new 'open' AI models, with the most capable one being competitive with OpenAI's o3-mini on at least one benchmark.
On Wednesday, Microsoft unveiled several new “open” AI models, with the most capable among them rivaling OpenAI’s o3-mini on a benchmark. The new permissively licensed models - Phi 4 mini reasoning, Phi 4 reasoning, and Phi 4 reasoning plus - fall under the category of “reasoning” models.
Advancements in AI Models
In the rapidly evolving world of artificial intelligence, breakthroughs continue to push the boundaries of what is achievable. While larger models garner a lot of attention, there is significant progress in creating more efficient and smaller systems. This is especially crucial for developers working on applications at the ‘edge’ or for use cases that demand lower latency and fewer computational resources.
Despite the progress made in large language models (LLMs), performance on reasoning-intensive tasks such as mathematical problem solving, algorithmic planning, or coding is still limited by factors like model size, training methodology, and inference-time capabilities. Models that excel in general NLP benchmarks often lack the capacity to create multi-step reasoning chains or analyze intermediate problem-solving stages.
New Compact AI Models by Microsoft
Microsoft introduced several advanced artificial intelligence systems with improved problem-solving capabilities. One of these versions effectively competes with larger counterparts in benchmark tests, showcasing enhanced performance.
To view factuality data please Upgrade to Premium. To view ownership data please Upgrade to Vantage.
