Phi-3 Vision 128K Instruct on Raspberry Pi 5 (8 GB)

Can the Raspberry Pi 5 (8 GB) run Phi-3 Vision 128K Instruct?

Yes, but slowly (tight fit)

Slow (tight fit). At Q3_K_M it needs about 6.4 GB of 6.4 GB usable, with an estimated speed of 5.4 tok/s.

Best quant: Q3_K_M
Generation speed: 5.4 tok/s
Prompt processing: 21.7 tok/s
First token (1K prompt): 47.2 s
Largest safe context: 8K

Best app to run it on the Raspberry Pi 5 (8 GB)

On Linux on ARM boards

Docker Model Runner

Command line + local API · runs on llama.cpp, vLLM

Handy if you already work in Docker.

docker model run hf.co/microsoft/Phi-3-vision-128k-instruct

How we choose: the files published for this model, your operating system and GPU, whether it fits fully in memory, and published speed tests (MLX vs llama.cpp on Apple silicon, vLLM vs llama.cpp vs Ollama on NVIDIA). Details.

Download Phi-3 Vision 128K Instruct

Original weights · Hugging Face

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

huggingface.co/microsoft/Phi-3-vision-128k-instruct (opens in a new tab)

No sign-in needed.

Links checked 1 day ago against Hugging Face, Ollama and GitHub.

How it’s rated

Quality 28Basic quality (D) · 28 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.: How capable the model itself is at Q3_K_M, from its editorial quality tier scaled by the quantization or precision. It’s the same on any hardware that runs this variant. Scale: A+ 85+ · A 70+ · B 55+ · C 40+ · D under 40.
Runs Slowtight fit: How well it runs on the Raspberry Pi 5 (8 GB): speed, capped by how well it fits in memory. Runs great: 30+ tokens/s (or near the image, video or decision speed target) · Runs well: 10+ tokens/s · Slow: 3+ tokens/s · Very slow: under 3 · Won’t run: doesn’t fit even with offload. A tight fit or RAM offload caps it at Runs well; heavy offload caps it at Slow.

Match score

To order lists, each result also gets a match score out of 100 from four parts. It favors the most capable model that still runs well.

Model qualityEditorial quality tier of the model, scaled by the quantization or precision used.	12.9 / 45
Memory fitFull marks when the model fits in the share of unified memory the GPU can use; fewer when it goes past that limit.	24.0 / 25
SpeedGeneration speed against a 40 tokens/s target for LLMs, or render time against a target for image and video.	5.5 / 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
Match score	57.4 / 100

The match score orders results; it is not shown as a grade. Quality and Runs are the two ratings in the tables.

Every quantization on the Raspberry Pi 5 (8 GB)

Quant	Memory needed	Available	Speed	Runs	Quality	Download
`Q2_K`	6.0 GB	6.4 GB	6.9 tok/s	Slowtight fit	23Basic quality (D) · 23 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 55/100 (quality, fit, speed and memory use) sets the ranking.	All files (opens in a new tab)
`Q3_K_M`	6.4 GB	6.4 GB	5.4 tok/s	Slowtight fit	28Basic quality (D) · 28 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.	All files (opens in a new tab)
`Q4_K_M`	6.9 GB	6.4 GB	3.6 tok/s	Slowover memory limit	32Basic quality (D) · 32 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 44/100 (quality, fit, speed and memory use) sets the ranking.	All files (opens in a new tab)
`Q5_K_M`	7.3 GB	6.4 GB	3.0 tok/s	Very slowfar over memory limit	32Basic quality (D) · 32 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 37/100 (quality, fit, speed and memory use) sets the ranking.	All files (opens in a new tab)
`Q6_K`	7.7 GB	6.4 GB	—	Won't run	33Basic quality (D) · 33 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 4/100 (quality, fit, speed and memory use) sets the ranking.	All files (opens in a new tab)
`Q8_0`	8.7 GB	6.4 GB	—	Won't run	34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 4/100 (quality, fit, speed and memory use) sets the ranking.	All files (opens in a new tab)
`F16`	12.6 GB	6.4 GB	—	Won't run	34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 4/100 (quality, fit, speed and memory use) sets the ranking.	Original weights (opens in a new tab)

Memory figures assume 8K of context. Longer conversations or documents grow the KV cache.

Other hardware that runs it

34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. Intel Arc B580 Q8_0 · 52.5 tok/s · Runs great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 3060 12GB Q8_0 · 54.2 tok/s · Runs great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4070 SUPER Q8_0 · 75.9 tok/s · Runs great
34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4070 Ti SUPER F16 · 46.6 tok/s · Runs great
34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4080 SUPER F16 · 51.1 tok/s · Runs great
34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 5070 Ti F16 · 62.2 tok/s · Runs great

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.