Leaderboard / Microsoft (Phi)

Phi-4-reasoning-plus

Name: Phi-4-reasoning-plus
Author: Microsoft (Phi)

Microsoft (Phi) · Phi-4 · released 2025-04-30 · data checked 2026-11-01 Needs verification

Phi-4-reasoning-plus is a 14.7B-parameter dense model from Microsoft (Phi) (United States), released 2025-04-30 under the MIT license. It scores 54.4 on the Small LM Index, #8 of 26 small models. It needs about 11.4 GB of memory at 4-bit with an 8K-token context.

Small LM Index

54.4

#8 of 26 scored small models

Likely 49.9–58.7 · from 4 of 6 core benchmarks

Parameters

14.7B

Dense

Price per 1M tokens

—

No hosted API on OpenRouter; self-host

Context window

32K

tokens

Benchmarks

Each bar shows this model's score. The tick marks the best small model on that benchmark. Sources: 4 reported by Microsoft (Phi). How we label sources.

MMLU-Pro Lab-reported Knowledge76.0

Class best: Qwen3.5 9B 82.5

GPQA Diamond Lab-reported Knowledge69.3

Class best: Qwen3.5 9B 81.7

AIME Lab-reported Math · AIME 202578.0

Class best: Apriel 1.6 15B Thinker 88.0

LiveCodeBench Lab-reported Coding53.1

Class best: Granite 4.2 8B 73.2

Not reported: estimated

What the index model expects from this model's other scores. Not a measurement.

Humanity's Last Exam≈ 8

SWE-bench Verified≈ 37

Specifications

Lab: Microsoft (Phi) (US · United States)
Architecture: Dense transformer
Total parameters: 14.7B
Active parameters: All (dense)
Size class: Small
Context window: 32K tokens
Input modalities: text
Reasoning: Reasoning (thinking)
License: MIT
Openness: Open weights
Memory to run: ~11.4 GB at 4-bit, 8K context (more sizes)
Hugging Face: 4.9K downloads (30 days) · 354 likes

Weights on Hugging Face ↗

Compare with

Phi-4-reasoning-plus vs Apriel 1.6 15B Thinker (14.9B, index 60.9)
Phi-4-reasoning-plus vs Ministral 3 14B (14B, index 60.0)
Phi-4-reasoning-plus vs Gemma 4 12B (11.9B, index 56.9)
Phi-4-reasoning-plus vs Phi-4 (14.7B, index 44.7)
Phi-4-reasoning-plus vs LLaDA2.2 mini (16.3B, index 29.7)
Phi-4-reasoning-plus vs Gemma 3 12B (12.2B, index 28.4)

Sources

https://huggingface.co/microsoft/Phi-4-reasoning-plus

Something wrong here? Report it.

Run it locally

Uses unsloth/Phi-4-reasoning-plus-GGUF (Q4_K_M). Swap the tag for another size: BF16, IQ4_NL, IQ4_XS, Q2_K, Q2_K_L, Q3_K_M, Q3_K_S, Q4_0….

Ollama

ollama run hf.co/unsloth/Phi-4-reasoning-plus-GGUF:Q4_K_M

llama.cpp

llama-server -hf unsloth/Phi-4-reasoning-plus-GGUF:Q4_K_M

Then open http://localhost:8080 for a chat window.

MLX (Mac)

pip install -U mlx-lm
mlx_lm.chat --model mlx-community/Phi-4-reasoning-plus-4bit

vLLM (GPU server)

vllm serve microsoft/Phi-4-reasoning-plus

Full-precision weights, for a server GPU.

Open in LM Studio

Memory needed: weights, context cache and runtime buffers.

Quantization	4K context	32K context
16-bit F16	31 GB	37 GB
8-bit Q8_0	17 GB	23 GB
4-bit Q4_K_M	11 GB	16 GB

Check it against your GPU, Mac or phone →