PhiBench
PhiBench is an internal benchmark designed to evaluate diverse skills and reasoning abilities of language models, covering a wide range of tasks including coding (debugging, extending incomplete code, explaining code snippets) and mathematics (identifying proof errors, generating related problems). Created by Microsoft's research team to address limitations of standard academic benchmarks and guide the development of the Phi-4 model.
Phi 4 Reasoning Plus from Microsoft currently leads the PhiBench leaderboard with a score of 0.742 across 3 evaluated AI models.
What PhiBench measures
PhiBench is a text benchmark that evaluates large language models on math, reasoning, and general tasks. LLM Stats tracks 3 models on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.7, with the leader reaching 0.7.
Compare leaders on the best AI for math, best AI for reasoning and best AI for general leaderboards.
Publication
- Paper
- Phi-4 Technical Report
- Authors
- Marah Abdin, Jyoti Aneja, Harkirat Behl, Sébastien Bubeck, and 23 others
- Published
- arXiv
- 2412.08905
Abstract
We present phi-4, a 14-billion parameter language model developed with a training recipe that is centrally focused on data quality. Unlike most language models, where pre-training is based primarily on organic data sources such as web content or code, phi-4 strategically incorporates synthetic data throughout the training process. While previous models in the Phi family largely distill the capabilities of a teacher model (specifically GPT-4), phi-4 substantially surpasses its teacher model on STEM-focused QA capabilities, giving evidence that our data-generation and post-training techniques go beyond distillation. Despite minimal changes to the phi-3 architecture, phi-4 achieves strong performance relative to its size -- especially on reasoning-focused benchmarks -- due to improved data, training curriculum, and innovations in the post-training scheme.
Phi 4 Reasoning Plus leads with 74.2%, followed by Phi 4 Reasoning at 70.6% and Phi 4 at 56.2%.
Progress Over Time
Interactive timeline showing model performance evolution on PhiBench
PhiBench Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 | Microsoft | 14B | — | — | ||
| 2 | Microsoft | 14B | — | — | ||
| 3 | Microsoft | 15B | — | — |
FAQ
Common questions about PhiBench.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions