HoML Logo

HoML
← Back to Models

Phi-3

The Phi-3 family are small, cost-effective language models from Microsoft that are powerful and outperform models of similar and larger sizes on various benchmarks. They are instruction-tuned and ready for use 'off-the-shelf'.

phi-3-mini-4k-instruct

3.8B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:mini-4k-instruct

Resource Requirements

Quantization Disk Space GPU Memory
FP16 7.6 GB 7.6 GB

phi-3-mini-4k-instruct-4bit

3.8B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:mini-4k-instruct-4bit

Resource Requirements

Quantization Disk Space GPU Memory
4-bit 1.8 GB 4 GB

phi-3-mini-128k-instruct

3.8B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:mini-128k-instruct

Resource Requirements

Quantization Disk Space GPU Memory
FP16 7.6 GB 7.6 GB

phi-3-mini-128k-instruct-4bit

3.8B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:mini-128k-instruct-4bit

Resource Requirements

Quantization Disk Space GPU Memory
4-bit 1.8 GB 4 GB

phi-3-small-8k-instruct

7B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:small-8k-instruct

Resource Requirements

Quantization Disk Space GPU Memory
FP16 14 GB 14 GB

phi-3-small-8k-instruct-4bit

7B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:small-8k-instruct-4bit

Resource Requirements

Quantization Disk Space GPU Memory
4-bit 3.5 GB 5 GB

phi-3-small-128k-instruct

7B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:small-128k-instruct

Resource Requirements

Quantization Disk Space GPU Memory
FP16 14 GB 14 GB

phi-3-small-128k-instruct-4bit

7B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:small-128k-instruct-4bit

Resource Requirements

Quantization Disk Space GPU Memory
4-bit 3.5 GB 5 GB

phi-3-medium-4k-instruct

14B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:medium-4k-instruct

Resource Requirements

Quantization Disk Space GPU Memory
FP16 28 GB 28 GB

phi-3-medium-4k-instruct-4bit

14B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:medium-4k-instruct-4bit

Resource Requirements

Quantization Disk Space GPU Memory
4-bit 7 GB 9 GB

phi-3-medium-128k-instruct

14B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:medium-128k-instruct

Resource Requirements

Quantization Disk Space GPU Memory
FP16 28 GB 28 GB

phi-3-medium-128k-instruct-4bit

14B parameters Instruction-Tuned

Pull this model

Use the following command with the HoML CLI:

homl pull phi-3:medium-128k-instruct-4bit

Resource Requirements

Quantization Disk Space GPU Memory
4-bit 7 GB 9 GB