unsloth / llama family / image-text-to-text

llama-4-maverick

meta/llama-4-maverick

GGUF Provisional scores Updated 18 Jun 2025

Run locally Smallest indexed GGUF / needs Ollama

ollama run hf.co/meta/llama-4-maverick/Llama-4-Maverick-17B-128E-Instruct-UD-TQ1_0.gguf

Benchmarks

Where it scores.

Methodology →

General Provisional

68.7

Coding Provisional

69.8

Math Provisional

69.8

Reasoning Provisional

64.7

Normalized 0–100 across indexed models.

Runnable variants

Pick a size.

27 indexed

Precision	Format	Size	Params	File
Q4 / smallest	GGUF	105.7 GB	—	Llama-4-Maverick-17B-128E-Instruct-UD-TQ1_0.gguf
IQ1_S	GGUF	120.8 GB	—	UD-IQ1_S/Llama-4-Maverick-17B-128E-Instruct-UD-IQ1_S-00001-of-00003.gguf
IQ1_M	GGUF	127.5 GB	—	UD-IQ1_M/Llama-4-Maverick-17B-128E-Instruct-UD-IQ1_M-00001-of-00003.gguf
IQ2_XXS	GGUF	135.2 GB	—	UD-IQ2_XXS/Llama-4-Maverick-17B-128E-Instruct-UD-IQ2_XXS-00001-of-00003.gguf
IQ2_M	GGUF	141.2 GB	—	UD-IQ2_M/Llama-4-Maverick-17B-128E-Instruct-UD-IQ2_M-00001-of-00003.gguf
Q2_K	GGUF	145.6 GB	—	Q2_K/Llama-4-Maverick-17B-128E-Instruct-Q2_K-00001-of-00003.gguf
Q2_K_L	GGUF	145.9 GB	—	Q2_K_L/Llama-4-Maverick-17B-128E-Instruct-Q2_K_L-00001-of-00003.gguf
Q2_K_XL	GGUF	152.7 GB	—	UD-Q2_K_XL/Llama-4-Maverick-17B-128E-Instruct-UD-Q2_K_XL-00001-of-00004.gguf
IQ3_XXS	GGUF	169.3 GB	—	UD-IQ3_XXS/Llama-4-Maverick-17B-128E-Instruct-UD-IQ3_XXS-00001-of-00004.gguf
Q3_K_S	GGUF	172.7 GB	—	Q3_K_S/Llama-4-Maverick-17B-128E-Instruct-Q3_K_S-00001-of-00004.gguf
Q3_K_XL	GGUF	179.6 GB	—	UD-Q3_K_XL/Llama-4-Maverick-17B-128E-Instruct-UD-Q3_K_XL-00001-of-00004.gguf
Q3_K_M	GGUF	191.1 GB	—	Q3_K_M/Llama-4-Maverick-17B-128E-Instruct-Q3_K_M-00001-of-00004.gguf
IQ4_XS	GGUF	214.3 GB	—	IQ4_XS/Llama-4-Maverick-17B-128E-Instruct-IQ4_XS-00001-of-00005.gguf
IQ4_NL	GGUF	225.8 GB	—	IQ4_NL/Llama-4-Maverick-17B-128E-Instruct-IQ4_NL-00001-of-00005.gguf
Q4_0	GGUF	226.8 GB	—	Q4_0/Llama-4-Maverick-17B-128E-Instruct-Q4_0-00001-of-00005.gguf
Q4_K_S	GGUF	227.8 GB	—	Q4_K_S/Llama-4-Maverick-17B-128E-Instruct-Q4_K_S-00001-of-00005.gguf
Q4_K_XL	GGUF	232.1 GB	—	UD-Q4_K_XL/Llama-4-Maverick-17B-128E-Instruct-UD-Q4_K_XL-00001-of-00005.gguf
Q4_K_M	GGUF	242.8 GB	—	Q4_K_M/Llama-4-Maverick-17B-128E-Instruct-Q4_K_M-00001-of-00005.gguf
Q4_1	GGUF	250.7 GB	—	Q4_1/Llama-4-Maverick-17B-128E-Instruct-Q4_1-00001-of-00006.gguf
Q5_K_S	GGUF	275.7 GB	—	Q5_K_S/Llama-4-Maverick-17B-128E-Instruct-Q5_K_S-00001-of-00006.gguf
Q5_K_M	GGUF	284.5 GB	—	Q5_K_M/Llama-4-Maverick-17B-128E-Instruct-Q5_K_M-00001-of-00006.gguf
Q5_K_XL	GGUF	287.0 GB	—	UD-Q5_K_XL/Llama-4-Maverick-17B-128E-Instruct-UD-Q5_K_XL-00001-of-00006.gguf
Q6_K	GGUF	328.8 GB	—	Q6_K/Llama-4-Maverick-17B-128E-Instruct-Q6_K-00001-of-00007.gguf
Q6_K_XL	GGUF	341.1 GB	—	UD-Q6_K_XL/Llama-4-Maverick-17B-128E-Instruct-UD-Q6_K_XL-00001-of-00008.gguf
Q8_0	GGUF	425.8 GB	—	Q8_0/Llama-4-Maverick-17B-128E-Instruct-Q8_0-00001-of-00009.gguf
Q8_K_XL	GGUF	460.0 GB	—	UD-Q8_K_XL/Llama-4-Maverick-17B-128E-Instruct-UD-Q8_K_XL-00001-of-00010.gguf
BF16	GGUF	801.5 GB	—	BF16/Llama-4-Maverick-17B-128E-Instruct-BF16-00001-of-00018.gguf

Sizes are file sizes, not peak runtime memory.

Raw benchmarks

The inputs.

3 records

Source	Task	Raw	Normalized
ZeroEval	GPQA	69.8	69.8
ZeroEval	MMMU	73.4	73.4
ZeroEval	MMMUPro	59.6	59.6

From the Hugging Face model card

About this model.

Read it on Hugging Face ↗