VOOZH about

URL: https://willitrunai.com/models/phi-3-mini-3.8b

โ‡ฑ Phi 3 Mini 3.8B VRAM Requirements โ€” GPU Compatibility


๐Ÿ‘ Microsoft
Microsoft

Phi 3 Mini 3.8B

Current
๐Ÿ‘ huggingface
HuggingFace๐Ÿ‘ ollama
Ollama
247.0KDownloads1.7KLikesApr 2024Released128K tokensContextMITLicense42 BasicQuality

Phi 3 Mini 3.8B (3.799999952316284B parameters) requires approximately 10.0 GB of VRAM with Q4_K_M quantization. For the best balance of quality and speed, we recommend hardware with at least 12 GB of VRAM.

Get started

โ€” copy & paste to run locally

Copy-paste commands to run Phi 3 Mini 3.8B on your machine.

Run

ollama run phi3:mini

Quick specs

Parameters3.8B
Architecturedense
Context128K tokens
Modalitytext
Min RAM1.5 GB
Rec. RAM2.3 GB (Q4_K_M)
LicenseMIT
FamilyPhi
โœ“ Chat

About this model

The Phi-3-Mini-128K-Instruct is a 3.8 billion-parameter, lightweight, state-of-the-art open model trained using the Phi-3 datasets. This dataset includes both synthetic data and filtered publicly available website data, with an emphasis on high-quality and reasoning-dense properties. The model belongs to the Phi-3 family with the Mini version in two variants 4K and 128K which is the context length (in tokens) that it can support.

  • โ€ขMemory/compute constrained environments
  • โ€ขLatency bound scenarios
  • โ€ขStrong reasoning (especially code, math and logic)

Related models

Your hardware

Detecting...

Quick picks

๐Ÿ‘ Intel
Best budgetB
Intel Arc B580 12GB~$249 โ€” 53 tok/s
๐Ÿ‘ NVIDIA
Best overallA
RTX 5070 Ti 16GB~$749 โ€” 72 tok/s

Best hardware

Top picks for Phi 3 Mini 3.8B

RTX 5070 Ti 16GBA
16 GB
RTX 5080 16GBA
16 GB
RTX 5060 Ti 16GBA
16 GB
RTX 4060 Ti 16GBA
16 GB
RTX 4070 Ti Super 16GBA
16 GB

Run this model

Phi 3 Mini 3.8B on RTX 5070 Ti 16GBPhi 3 Mini 3.8B on RTX 5080 16GBPhi 3 Mini 3.8B on RTX 5060 Ti 16GB

Quantization options

VRAM estimates by quant level

No hardware detected โ€” fit column shows raw VRAM estimates

QuantBitsVRAMQualityFit
Q2_K
2
1.5 GB
Lowโ€”
Q3_K_S
3
1.9 GB
Lowโ€”
NVFP4
4
2.1 GB
Mediumโ€”
Q4_K_M
4
2.3 GB
Mediumโ€”
Q5_K_M
5
2.7 GB
Highโ€”
Q6_K
6
3.1 GB
Highโ€”
Q8_0
8
4.1 GB
Very Highโ€”
F16
16
7.8 GB
Maximumโ€”

Quality benchmarks

Phi 3 Mini 3.8B benchmark scores

Benchmark verified

Coding

SWE-bench Verifiedโ€”
HumanEval+60.4%
Aider Polyglotโ€”
LiveCodeBenchโ€”

Reasoning

MMLU-Pro69.7%
GPQA Diamond29.7%
MATH-50014.0%
ARC Challenge85.5%

General

Chatbot Arenaโ€”
IFEval59.8%

Source: official ยท 2024-04-22

Hardware compatibility

Fit estimates across all hardware

Open calculator

Computing compatibility...

Memory breakdown

Reference: RTX 2060 6GB

Weights2.3 GB
KV Cache5.9 GB
Runtime1.2 GB
Headroom0.6 GB

Frequently asked questions

FAQ โ€” Phi 3 Mini 3.8B

See also

Quantization GuideScoring MethodologyVRAM Calculator