Kimi Linear 48B-A3B on NVIDIA Jetson AGX Thor (128 GB)

Can the NVIDIA Jetson AGX Thor (128 GB) run Kimi Linear 48B-A3B?

Yes, it runs great

Runs great. At Q8_0 it needs about 61.9 GB of 115 GB usable, with an estimated speed of 46.8 tok/s.

Best quant: Q8_0
Generation speed: 46.8 tok/s
Prompt processing: 1,920 tok/s
First token (1K prompt): 0.5 s
Largest safe context: 1,024K

Best app to run it on the NVIDIA Jetson AGX Thor (128 GB)

On Linux on ARM boards

vLLM

Inference server · runs on 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.

vllm serve moonshotai/Kimi-Linear-48B-A3B-Instruct

LM Studio Point-and-click: search, download and chat in one window.
llama.cpp Usually the first to support new model architectures.

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 Kimi Linear 48B-A3B

Recommended for the NVIDIA Jetson AGX Thor (128 GB): Q8_0. Every other quantization has its own download link in the table below.

Quantized files · GGUF

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

Download the Q8_0 files (opens in a new tab)

huggingface.co/bartowski/moonshotai_Kimi-Linear-48B-A3B-Instruct-GGUF (opens in a new tab)

ollama run hf.co/bartowski/moonshotai_Kimi-Linear-48B-A3B-Instruct-GGUF

Original weights · Hugging Face

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

huggingface.co/moonshotai/Kimi-Linear-48B-A3B-Instruct (opens in a new tab)

No sign-in needed.

Official repo · GitHub

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

github.com/MoonshotAI/Kimi-Linear (opens in a new tab)

git clone https://github.com/MoonshotAI/Kimi-Linear

Links checked 3 days ago against Hugging Face, Ollama and GitHub.

How it’s rated

Quality 74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.: How capable the model itself is at Q8_0, 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 NVIDIA Jetson AGX Thor (128 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.	33.3 / 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.	12.4 / 15
Match score	85.7 / 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 NVIDIA Jetson AGX Thor (128 GB)

Quant	Memory needed	Available	Speed	Runs	Quality	Download
`Q2_K`	29.5 GB	115 GB	117 tok/s	Runs great	51Fair quality (C) · 51 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q3_K_M`	34.0 GB	115 GB	98.4 tok/s	Runs great	62Good quality (B) · 62 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q4_K_M`	39.8 GB	115 GB	81.2 tok/s	Runs great	70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q5_K_M`	44.8 GB	115 GB	69.8 tok/s	Runs great	71Great quality (A) · 71 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q6_K`	50.2 GB	115 GB	60.6 tok/s	Runs great	73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q8_0`	61.9 GB	115 GB	46.8 tok/s	Runs great	74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab)
`F16`	108 GB	115 GB	23.5 tok/s	Runs welltight fit	74Great quality (A) · 74 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	GGUF · folder (opens in a new tab) 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

74Great quality (A) · 74 of 100: how capable the model is at F16, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking. Apple M2 Ultra (192 GB) F16 · 51.3 tok/s · Runs great
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. Apple M4 Max (128 GB) Q8_0 · 71.4 tok/s · Runs great
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. AMD Ryzen AI Max+ 395 (128 GB) Q8_0 · 39.1 tok/s · Runs great
71Great quality (A) · 71 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. Apple M1 Max (64 GB) Q5_K_M · 82.7 tok/s · Runs great
71Great quality (A) · 71 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. Apple M4 Max (64 GB) Q5_K_M · 102 tok/s · Runs great
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA RTX PRO 6000 Blackwell Q8_0 · 154 tok/s · Runs great

Quick answers

How much memory does Kimi Linear 48B-A3B need on the NVIDIA Jetson AGX Thor 128GB?

At Q8_0 it needs about 61.9 GB. The NVIDIA Jetson AGX Thor 128GB has 115 GB usable.

How fast is Kimi Linear 48B-A3B on the NVIDIA Jetson AGX Thor 128GB?

About 46.8 tok/s at Q8_0, an estimate from the hardware’s memory bandwidth and the size of the model.

Which quantization of Kimi Linear 48B-A3B should I use on the NVIDIA Jetson AGX Thor 128GB?

Q8_0. 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.