LLMs · LLaVA

LLaVA 1.6 34B

Name: LLaVA 1.6 34B
Author: LLaVA

Parameters: 35B
Architecture: Dense
Max context: 4K tokens
License: apache-2.0
Released: 2024-03

How to run it

Quickest · Ollama

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

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

ollama run llava:34b

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 2 days ago against Hugging Face, Ollama and GitHub.

Apps that run LLaVA 1.6 34B

Our pick for each kind of computer, assuming the model fits in its memory. Pick your own GPU or device in the Model Finder for an exact answer.

Mac (Apple silicon)

Ollama

One command with Ollama’s own build of this model.

Also good:

Windows PC, NVIDIA GPU

Ollama

One command with Ollama’s own build of this model.

Also good:

Windows PC, AMD Radeon

Ollama

One command with Ollama’s own build of this model.

Also good:

Linux, NVIDIA GPU

Ollama

One command with Ollama’s own build of this model.

Also good:

Linux server, 80 GB+ GPUs

vLLM

The full model fits in GPU memory at 8-bit, so it can batch many requests: best for serving a team or an app.

Also good:

App	Type	Opens it as	Systems	Engine	Start
Ollama	CLI + API	Ollama library	macOSWindowsLinux	llama.cpp, MLX	`ollama run llava:34b`
vLLM	Server	Original weights	WSLWindowsLinux	vLLM	`vllm serve llava-hf/llava-v1.6-34b-hf`
SGLang	Server	Original weights	WSLWindowsLinux	SGLang	`python3 -m sglang.launch_server --model-path llava-hf/llava-v1.6-34b-hf`
Docker Model Runner	CLI + API	Original weights	macOSWindowsLinux	llama.cpp, vLLM	`docker model run hf.co/llava-hf/llava-v1.6-34b-hf`
Text Generation WebUI	Desktop app	Original weights	macOSWindowsLinux	llama.cpp, Transformers, ExLlamaV3	Get Text Generation WebUI

Which apps can open a model depends on the files published for it: GGUF for llama.cpp-based apps, MLX for Apple’s engine, the original weights for vLLM and SGLang. Compare every app.

Memory needed

Weights plus KV cache and runtime overhead. Leave a little extra free for the operating system and other apps. Try another context length.

Quantization	Quality	Weights	Total at 4K context	Download
`Q2_K`	Low	13.9 GB	15.5 GB
`Q3_K_M`	Low	17.0 GB	18.6 GB
`Q4_K_M`	Good	21.2 GB	22.8 GB
`Q5_K_M`	High	24.7 GB	26.3 GB
`Q6_K`	High	28.5 GB	30.1 GB
`Q8_0`	Very high	36.8 GB	38.4 GB
`F16`	Full	69.5 GB	71.1 GB	Original weights (opens in a new tab)

Best hardware for LLaVA 1.6 34B

The least expensive GPU in our catalog that runs it without offloading is the AMD Radeon RX 7900 XT (launch price $899) at Q3_K_M.

Quality	Hardware	Quant	Memory	Speed	Runs
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.	NVIDIA GeForce RTX 5090	`Q4_K_M`	24.4 GB / 32.0 GB	63.3 tok/s	Runs great
36Basic quality (D) · 36 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.	NVIDIA GeForce RTX 3090	`Q3_K_M`	19.8 GB / 24.0 GB	42.7 tok/s	Runs great
36Basic quality (D) · 36 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.	NVIDIA GeForce RTX 4090	`Q3_K_M`	19.8 GB / 24.0 GB	46.0 tok/s	Runs great
36Basic quality (D) · 36 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.	AMD Radeon RX 7900 XTX	`Q3_K_M`	19.8 GB / 24.0 GB	37.8 tok/s	Runs great
43Fair quality (C) · 43 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.	NVIDIA RTX PRO 6000 Blackwell	`Q6_K`	34.9 GB / 96.0 GB	43.9 tok/s	Runs great
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.	Apple M4 Max (64 GB)	`Q4_K_M`	26.6 GB / 48.0 GB	16.6 tok/s	Runs well
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.	Apple M1 Max (64 GB)	`Q4_K_M`	26.6 GB / 48.0 GB	12.2 tok/s	Runs well
36Basic quality (D) · 36 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.	Apple M4 Pro (48 GB)	`Q3_K_M`	21.5 GB / 36.0 GB	10.7 tok/s	Runs well
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.	Apple M2 Ultra (192 GB)	`Q4_K_M`	34.3 GB / 144 GB	24.3 tok/s	Runs well
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.	Apple M4 Max (128 GB)	`Q4_K_M`	30.5 GB / 96.0 GB	16.6 tok/s	Runs well

Check it on your own hardware

Run it on a cluster

Devices that can’t run LLaVA 1.6 34B alone can run it together. These example clusters do:

Four Raspberry Pi 5s (16 GB)0.3 tok/s at Q6_K with llama.cpp RPC

Try it with your own devices

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.