MR
Microsoft Research
Language model · United States ·Dec 2024

Phi-4

Language Open (Restricted)
14B
Parameters
1
Benchmarks

About Phi-4

Phi-4 is an AI model developed by Microsoft Research, in the language category, released in 2024, made available as an open-weights model with 14B params.

On this page you'll find Phi-4's full specifications, including 1 benchmark result. Review provider pricing and benchmark scores below, or compare Phi-4 head-to-head with other language models.

→ More models by Microsoft Research → All language models → AI benchmarks leaderboard

Links

Announcement ↗ Hugging Face ↗

Run Phi-4 locally — what to buy

Share this build ↗

Sized to what Phi-4 actually needs (~8k context at Q4), not to the biggest GPU: Good = cheapest that runs it, Better = best value, Best = most headroom. Speed is a hardware estimate; anything past ~120 tok/s is shown as instant.

Good
Intel Arc B580 12GB
Consumer GPU · ~$249
~41.9 tok/s
Q4 · fits full context · up to 8k ctx
Runs with: ExLlamaV2, KoboldCpp, LM Studio
See full estimate →
Better
Intel Arc A770 16GB
Consumer GPU · ~$329
~51.4 tok/s
Q4 · fits full context · up to 8k ctx
Runs with: ExLlamaV2, KoboldCpp, LM Studio
See full estimate →
Best
NVIDIA GeForce RTX 4090
Consumer GPU · ~$1,599
~111.9 tok/s
Q4 · fits full context · up to 8k ctx
Runs with: ExLlamaV2, KoboldCpp, LM Studio
See full estimate →

Software support

·ExLlamaV2 ·KoboldCpp ·LM Studio ·MLX ·Ollama ·SGLang ·TensorRT-LLM ·llama.cpp ·vLLM ?KTransformers ?text-generation-webui

✓ measured · · compatible · — not supported. Informational only — speed is hardware-based.

Consumer GPU
20.6–120+ tok/s
27 options →
Apple
31.4–83.7 tok/s
17 options →
Datacenter GPU
84.9–120+ tok/s
13 options →

Community

Would you run Phi-4 again?
No votes yet — be the first.
Did it run? — community reports by hardware

No reports yet. Own this model on some hardware? Be the first to confirm it runs.

Discussion

No comments yet. Start the discussion.

Phi-4 benchmark scores▶
Compare on the benchmark table →
Benchmark Category Score Variant
AIME 2025 math 17% baseline-cited-phi4r

More models like Phi-4

AI
DMind-1
 AI
COLLECTIVE-1
 AI
Cogito 1 series
 AI
Euler (Euler大模型)
· 2.5B
 AI
SmolLM-1.7B
· 1.7B
 AI
Llama-3-70B-Synthia-v3.5
· 70B

Frequently asked questions

What is Phi-4? ▶

Phi-4 is an AI model developed by Microsoft Research, in the language category, released in 2024. It is tracked on GenAIList with its specifications, benchmark scores and provider pricing.

Who created Phi-4? ▶

Phi-4 was developed by Microsoft Research and released in 2024.

Is Phi-4 open source or proprietary? ▶

Phi-4 ships as an open-weights model: you can download and self-host the weights, though the licence may place some restrictions on use.

How much does Phi-4 cost? ▶

Pricing for Phi-4 depends on the provider. See the providers table on this page for the latest API rates.

How does Phi-4 perform on benchmarks? ▶

Phi-4 is benchmarked across 1 evaluation on GenAIList, including AIME 2025 (17%). See the full benchmark table below and compare it with other models.

No videos yet. Know a good one? Add it below.

Know a great video about Phi-4?