DeepSeek Coder V2 Lite (16B) on Apple M3 Ultra (256 GB)

Can the Apple M3 Ultra (256 GB) run DeepSeek Coder V2 Lite (16B)?

Yes, it runs great

Runs great. At F16 it needs about 47.6 GB of 192 GB usable, with an estimated speed of 61.8 tok/s.

Best quant: F16
Generation speed: 61.8 tok/s
Prompt processing: 7,337 tok/s
First token (1K prompt): 0.3 s
Largest safe context: 160K

Best app to run it on the Apple M3 Ultra (256 GB)

On macOS

LM Studio

Desktop app · runs on llama.cpp, MLX · use the MLX build

Runs the MLX build, typically 35–90% faster than GGUF for a model with 2.4B active parameters on Apple silicon.
Point-and-click: search, download and chat in one window.

Open DeepSeek Coder V2 Lite (16B) in LM Studio

MLX LM Runs the MLX build, typically 35–90% faster than GGUF for a model with 2.4B active parameters on Apple silicon.
Ollama One command with Ollama’s own build of this model.

Want an agent? DeepSeek Coder V2 Lite (16B) supports tool use, so you can pair a llama.cpp or MLX LM server with one of these. Apps that run on this computer come first.

Coding agents

Personal assistants

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 DeepSeek Coder V2 Lite (16B)

Recommended for the Apple M3 Ultra (256 GB): F16. 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.

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

ollama run deepseek-coder-v2:16b-lite-instruct-fp16

Quantized files · GGUF

Works in LM Studio, llama.cpp, Jan and Ollama. No sign-in needed.

huggingface.co/lmstudio-community/DeepSeek-Coder-V2-Lite-Instruct-GGUF (opens in a new tab)

ollama run hf.co/lmstudio-community/DeepSeek-Coder-V2-Lite-Instruct-GGUF

Original weights · Hugging Face

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

huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct (opens in a new tab)

No sign-in needed.

Official repo · GitHub

Reference code, inference scripts and docs from DeepSeek. The weights themselves download from Hugging Face.

github.com/deepseek-ai/DeepSeek-Coder-V2 (opens in a new tab)

git clone https://github.com/deepseek-ai/DeepSeek-Coder-V2

Links checked 2 days ago against Hugging Face, Ollama and GitHub.

How it’s rated

Quality 52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.: How capable the model itself is at F16, 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 Runs great: How well it runs on the Apple M3 Ultra (256 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.	23.4 / 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.	25.0 / 25
SpeedGeneration speed against a 40 tokens/s target for LLMs, or render time against a target for image and video.	15.0 / 15
Memory useRewards using the memory you have, so a bigger, better variant beats a tiny one that leaves most memory idle.	5.7 / 15
Match score	69.1 / 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 Apple M3 Ultra (256 GB)

Quant	Memory needed	Available	Speed	Runs	Quality	Download
`Q2_K`	22.5 GB	192 GB	208 tok/s	Runs great	36Basic quality (D) · 36 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.
`Q3_K_M`	23.9 GB	192 GB	186 tok/s	Runs great	43Fair quality (C) · 43 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (Q3_K_L) (opens in a new tab)
`Q4_K_M`	25.8 GB	192 GB	163 tok/s	Runs great	49Fair quality (C) · 49 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q5_K_M`	27.4 GB	192 GB	146 tok/s	Runs great	50Fair quality (C) · 50 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q6_K`	29.1 GB	192 GB	132 tok/s	Runs great	51Fair quality (C) · 51 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q8_0`	32.9 GB	192 GB	108 tok/s	Runs great	52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`F16`	47.6 GB	192 GB	61.8 tok/s	Runs great	52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 69/100 (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

52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. AMD Radeon RX 7900 XTX Q8_0 · 118 tok/s · Runs great
52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Apple M4 Max (64 GB) F16 · 45.3 tok/s · Runs great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 3090 Q8_0 · 126 tok/s · Runs great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4090 Q8_0 · 131 tok/s · Runs great
50Fair quality (C) · 50 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking. AMD Radeon RX 9070 XT Q5_K_M · 128 tok/s · Runs great
50Fair quality (C) · 50 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking. Apple M4 Pro (24 GB) Q5_K_M · 73.7 tok/s · Runs great

Quick answers

How much memory does DeepSeek Coder V2 Lite (16B) need on the M3 Ultra 256GB?

At F16 it needs about 47.6 GB. The M3 Ultra 256GB has 192 GB usable.

How fast is DeepSeek Coder V2 Lite (16B) on the M3 Ultra 256GB?

About 61.8 tok/s at F16, an estimate from the hardware’s memory bandwidth and the size of the model.

Which quantization of DeepSeek Coder V2 Lite (16B) should I use on the M3 Ultra 256GB?

F16. Of the versions that fit, it is the one that runs best; when two run equally well, the check prefers the higher-rated one.

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.