phi-4 reasoning models

[ysjprojects]

Models added:
Phi-4-mini-reasoning | 3.8B
Phi-4-reasoning | 14B
Phi-4-reasoning-plus | 14B

Relevant links:
https://arxiv.org/abs/2504.21233 (mini-reasoning)
https://arxiv.org/abs/2504.21318 (reasoning)
https://azure.microsoft.com/en-us/blog/one-year-of-phi-small-language-models-making-big-leaps-in-ai/

Highlight:

Despite their significantly smaller size, both models achieve better performance than OpenAI o1-mini and DeepSeek-R1-Distill-Llama-70B at most benchmarks, including mathematical reasoning and Ph.D. level science questions. They achieve performance better than the full DeepSeek-R1 model (with 671-billion parameters) on the AIME 2025 test, the 2025 qualifier for the USA Math Olympiad.