Phi-4 Reasoning Plus AI Evaluation | LayerLens - Benchmark Results & Performance

Model

Phi-4 Reasoning Plus by Microsoft

text->text

Free credits included. No credit card required.

Phi-4-reasoning-plus is an enhanced 14B parameter model from Microsoft, fine-tuned from Phi-4 with additional reinforcement learning to boost accuracy on math, science, and code reasoning tasks. It uses the same dense decoder-only transformer architecture as Phi-4, but generates longer, more comprehensive outputs structured into a step-by-step reasoning trace and final answer. Designed for English-only applications, it is well-suited for structured reasoning workflows where output quality takes priority over response speed.

Evaluations

across 18 benchmarks

Latency

—

Context length

33k

tokens

Cost

$0.07 · $0.35

input · output per 1M tokens

Key takeaways

Phi-4 Reasoning Plus is a 14B parameter, English-only Transformer model designed for structured reasoning tasks in math, science, and code, prioritizing output quality and comprehensive step-by-step explanations.
The model was fine-tuned from Phi-4 with additional reinforcement learning.
It demonstrates strong capabilities in basic scientific concepts, achieving 100% accuracy on the AI2 Reasoning Challenge - Easy dataset, consistently following instructions and providing correct answers.
A significant limitation is its complete failure (0% accuracy) on the SWE-bench Lite dataset, indicating a profound inability to handle complex, real-world software engineering issues or provide helpful responses.
Known limitations include potential minor formatting issues and struggles with software development workflows, suggesting a need for further training or fine-tuning to address complex scientific computing scenarios and improve its action and observation chains for problem-solving.

Technical specifications

Context length: 33k tokens
Max output: —
Architecture: Transformer
Modalities: text->text
Parameters: ~13B
Open weights: No
Licence: Proprietary
Regions: USA
Vendor: Microsoft
Provider model key: microsoft/phi-4-reasoning-plus