LLM Finder

Match local LLMs to your GPU. Filter by VRAM + use case, then copy the download command.

GPU VRAM

8 GB 12 GB 16 GB 24 GB 48 GB 80 GB

Use Case

all coding chat research creative math

21 models fit 48GB for math

Data powered by llmfit ↗

Qwen2.5-72B-Instruct

72Bq3_K_M • 38.4GB

Family: Qwen

Estimated speed: 37-47 tok/s

Context: 33k

License: Apache-2.0

chatcodingresearchmathcreative

ollama pull qwen2.5:72b

Llama-3.3-70B-Instruct

70Bq4_K_M • 47.4GB

Family: Llama

Estimated speed: 17-27 tok/s

Context: 131k

License: Llama 3.3 Community

chatcodingresearchcreativemath

ollama pull llama3.3:70b

Qwen2.5-Coder-32B-Instruct

32Bq8_0 • 39.6GB

Family: Qwen

Estimated speed: 35-45 tok/s

Context: 33k

License: Apache-2.0

codingchatmathresearch

ollama pull qwen2.5-coder:32b

Llama-3.1-70B-Instruct

70Bq4_K_M • 47.4GB

Family: Llama

Estimated speed: 18-28 tok/s

Context: 131k

License: Llama 3.1 Community

chatresearchcreativemath

ollama pull llama3.1:70b

DeepSeek-R1-Distill-Llama-70B

70Bq4_K_M • 47.4GB

Family: DeepSeek

Estimated speed: 18-28 tok/s

Context: 33k

License: MIT

mathresearchcodingchat

ollama pull deepseek-r1:70b

Yi-1.5-34B-Chat

34Bq8_0 • 41.9GB

Family: Yi

Estimated speed: 31-41 tok/s

Context: 33k

License: Apache-2.0

chatresearchcreativemath

ollama pull yi:34b

Qwen2.5-32B-Instruct

32Bq8_0 • 39.6GB

Family: Qwen

Estimated speed: 37-47 tok/s

Context: 33k

License: Apache-2.0

chatcodingresearchmathcreative

ollama pull qwen2.5:32b

Phi-4-14B-Instruct

14Bfp16 • 33.3GB

Family: Phi

Estimated speed: 52-62 tok/s

Context: 128k

License: MIT

codingmathresearchchat

ollama pull phi4:14b

Phi-3-medium-128k-instruct

14Bfp16 • 33.3GB

Family: Phi

Estimated speed: 53-63 tok/s

Context: 131k

License: MIT

codingchatresearchmath

ollama pull phi3:medium

DeepSeek-R1-Distill-Llama-8B

8Bfp16 • 20.1GB

Family: DeepSeek

Estimated speed: 86-96 tok/s

Context: 33k

License: MIT

mathresearchchat

ollama pull deepseek-r1:8b

Qwen2.5-Coder-14B-Instruct

14Bfp16 • 33.3GB

Family: Qwen

Estimated speed: 54-64 tok/s

Context: 33k

License: Apache-2.0

codingchatmath

ollama pull qwen2.5-coder:14b

Qwen2.5-14B-Instruct

14Bfp16 • 33.3GB

Family: Qwen

Estimated speed: 54-64 tok/s

Context: 33k

License: Apache-2.0

chatcodingresearchmath

ollama pull qwen2.5:14b

Mixtral-8x22B-Instruct

141B MoEq3_K_M • 47.8GB

Family: Mistral

Estimated speed: 22-32 tok/s

Context: 66k

License: Apache-2.0

chatcodingresearchcreativemath

ollama pull mixtral:8x22b

Gemma-2-27B-Instruct

27Bq8_0 • 33.7GB

Family: Gemma

Estimated speed: 54-64 tok/s

Context: 8k

License: Gemma Terms

chatcodingresearchcreativemath

ollama pull gemma2:27b

DeepSeek-R1-Distill-Qwen-14B

14Bfp16 • 33.3GB

Family: DeepSeek

Estimated speed: 56-66 tok/s

Context: 33k

License: MIT

mathresearchcoding

ollama pull deepseek-r1:14b

DeepSeek-Coder-33B-Instruct

33Bq8_0 • 40.7GB

Family: DeepSeek

Estimated speed: 40-50 tok/s

Context: 16k

License: DeepSeek License

codingchatmath

ollama pull deepseek-coder:33b

Qwen2.5-7B-Instruct

7Bfp16 • 17.9GB

Family: Qwen

Estimated speed: 95-105 tok/s

Context: 33k

License: Apache-2.0

chatcodingresearchmath

ollama pull qwen2.5:7b

CodeLlama-34B-Instruct

34Bq8_0 • 41.9GB

Family: CodeLlama

Estimated speed: 40-50 tok/s

Context: 16k

License: Llama 2 Community

codingmath

ollama pull codellama:34b

DeepSeek-R1-Distill-Qwen-7B

7Bfp16 • 17.9GB

Family: DeepSeek

Estimated speed: 98-108 tok/s

Context: 33k

License: MIT

mathresearchcoding

ollama pull deepseek-r1:7b

StarCoder2-15B-Instruct

15Bfp16 • 35.5GB

Family: StarCoder

Estimated speed: 58-68 tok/s

Context: 16k

License: OpenRAIL-M

codingmath

ollama pull starcoder2:15b

CodeLlama-13B-Instruct

13Bfp16 • 31.1GB

Family: CodeLlama

Estimated speed: 71-81 tok/s

Context: 16k

License: Llama 2 Community

codingmath

ollama pull codellama:13b