LLaVA 1.6 34B on AMD Radeon RX 9060 XT 8GB

Can the AMD Radeon RX 9060 XT 8GB run LLaVA 1.6 34B?

41 pts

Barely: most of the model spills into system RAM

Heavy offload. At Q4_K_M it needs about 23.2 GB of 8.0 GB usable, with an estimated speed of 3.0 tok/s.

Best quant: Q4_K_M
Generation speed: 3.0 tok/s
Prompt processing: 455 tok/s
First token (1K prompt): 2.3 s
Largest safe context: —

Download LLaVA 1.6 34B

Recommended for the AMD Radeon RX 9060 XT 8GB: Q4_K_M. Every other quantization has its own download link in the table below.

Quickest · Ollama

One command downloads a ready-to-run version and starts a chat.

ollama run llava:34b-v1.6-q4_K_M

Get Ollama at ollama.com/download (opens in a new tab) · model page (opens in a new tab)

Original weights · Hugging Face

Full-precision files from LLaVA, for fine-tuning, vLLM, Transformers.

huggingface.co/llava-hf/llava-v1.6-34b-hf (opens in a new tab)

No sign-in needed.

Links checked 7 days ago against Hugging Face, Ollama and GitHub.

How the D grade is scored

Every result gets up to 100 points from four parts. The points decide the letter grade.

Model qualityEditorial quality tier of the model, scaled by the quantization or precision used.	18.8 / 45
Memory fitFull marks when everything sits in GPU or unified memory; fewer when part spills to system RAM.	4.0 / 25
SpeedGeneration speed against a 40 tokens/s target for LLMs, or render time against a target for image and video.	4.1 / 15
Memory useRewards using the memory you have, so a bigger, better variant beats a tiny one that leaves most memory idle.	15.0 / 15
Total	41.9 / 100	41 ptsPoor (D) · 41 of 100 points: Model quality 18.8/45, Memory fit 4.0/25, Speed 4.1/15, Memory use 15.0/15

Grade scale: S 85+ · A 72+ · B 58+ · C 42+ · D 25+ · F under 25 points.

Every quantization on the AMD Radeon RX 9060 XT 8GB

Quant	Memory needed	Available	Speed	Fit	Grade · points	Download
`Q2_K`	15.9 GB	8.0 GB	5.7 tok/s	Heavy offload	38 ptsPoor (D) · 38 of 100 points: Model quality 13.9/45, Memory fit 4.0/25, Speed 5.7/15, Memory use 15.0/15
`Q3_K_M`	19.0 GB	8.0 GB	4.2 tok/s	Heavy offload	40 ptsPoor (D) · 40 of 100 points: Model quality 16.6/45, Memory fit 4.0/25, Speed 4.8/15, Memory use 15.0/15
`Q4_K_M`	23.2 GB	8.0 GB	3.0 tok/s	Heavy offload	41 ptsPoor (D) · 41 of 100 points: Model quality 18.8/45, Memory fit 4.0/25, Speed 4.1/15, Memory use 15.0/15
`Q5_K_M`	26.7 GB	8.0 GB	—	Won't run	5 ptsWon't run (F) · 5 of 100 points: Model quality 5.8/45, Memory fit 0.0/25, Speed 0.0/15, Memory use 0.0/15
`Q6_K`	30.5 GB	8.0 GB	—	Won't run	5 ptsWon't run (F) · 5 of 100 points: Model quality 5.9/45, Memory fit 0.0/25, Speed 0.0/15, Memory use 0.0/15
`Q8_0`	38.8 GB	8.0 GB	—	Won't run	5 ptsWon't run (F) · 5 of 100 points: Model quality 5.9/45, Memory fit 0.0/25, Speed 0.0/15, Memory use 0.0/15
`F16`	71.5 GB	8.0 GB	—	Won't run	5 ptsWon't run (F) · 5 of 100 points: Model quality 5.9/45, Memory fit 0.0/25, Speed 0.0/15, Memory use 0.0/15	Original weights (opens in a new tab)

Memory figures assume 4K of context. Longer conversations or documents grow the KV cache.

Other hardware that runs it

73 ptsGreat (A) · 73 of 100 points: Model quality 18.8/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 15.0/15 NVIDIA GeForce RTX 5090 Q4_K_M · 63.3 tok/s
71 ptsGood (B) · 71 of 100 points: Model quality 16.6/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 15.0/15 NVIDIA GeForce RTX 3090 Q3_K_M · 42.7 tok/s
71 ptsGood (B) · 71 of 100 points: Model quality 16.6/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 15.0/15 NVIDIA GeForce RTX 4090 Q3_K_M · 46.0 tok/s
71 ptsGood (B) · 71 of 100 points: Model quality 16.6/45, Memory fit 25.0/25, Speed 14.6/15, Memory use 15.0/15 AMD Radeon RX 7900 XTX Q3_K_M · 37.8 tok/s
67 ptsGood (B) · 67 of 100 points: Model quality 19.6/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 8.4/15 NVIDIA RTX PRO 6000 Blackwell Q6_K · 43.9 tok/s
66 ptsGood (B) · 66 of 100 points: Model quality 18.8/45, Memory fit 25.0/25, Speed 9.7/15, Memory use 12.8/15 Apple M4 Max (64 GB) Q4_K_M · 16.6 tok/s

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.