FitMyLLM
NVIDIA· VOLTA

NVIDIA Jetson AGX Xavier 16 GB

Running LLMs on the Jetson AGX Xavier 16 GB — the long read: which models fit at which quantisation, and the settings worth changing. · Or what a budget buys

VRAM
16 GB
MID-RANGE
BANDWIDTH
136
GB/S
MODELS Q4
262/449
58%
7B Q4 SPEED
~16
GOOD
▸ MODEL COVERAGE @ Q458% OF ALL
▸ ESTIMATED SPEED· BY MODEL SIZE @ Q4

Average speeds at Q4 quantization. Actual performance varies by model architecture and context length.

3B
~36
TOK/S
7B
~16
TOK/S
14B
~8
TOK/S
32B
18.0GB NEEDED
70B
39.4GB NEEDED
▸ MEASURED RIG REPORTS

We rent the machine and time every model on it: decode, VRAM peak, concurrency, watts and cost per million tokens. Yours may already be one of them — and two are free to read in full.

SEE THE REPORTS
▸ SPECIFICATIONS
VRAM
16 GB
BANDWIDTH
136 GB/s
FP16 COMPUTE
2.8 TFLOPS
TDP
30W
MEMORY
LPDDR4X
ARCHITECTURE
Volta
CUDA CORES
512
TENSOR CORES
64
PCIE
Gen 4 x4
93
FAST MODELS · >30 TOK/S
Real-time chat speed
226
USABLE · >10 TOK/S
Comfortable for all tasks
262
TOTAL COMPATIBLE
Fit in VRAM at Q4
▸ DON’T WANT TO BUY?

Test Jetson AGX Xavier 16 GB (or anything bigger) without committing. Pay by the second, cancel anytime.

Spin up in ~60s. Pay by the second. Cancel anytime.

Some links are affiliate links — we may earn a small commission at no extra cost to you. This helps keep FitMyLLM free and independent.

▸ COMPATIBLE MODELS· 262
S
Falcon-H1R Tiny 90M0.09B
FALCON·256K CTX· CHAT· REASONING
1209
TOK/S · 3% VRAM
S
SmolLM2 135M0.135B
SMOLLM·2K CTX· CHAT
806
TOK/S · 4% VRAM
S
nomic-embed-text-v1.5 100M0.14B
EMBEDDING·8K CTX· CHAT
777
TOK/S · 4% VRAM
S
GPT-2 124M0.14B
GPT2·1K CTX· CHAT
777
TOK/S · 4% VRAM
S
SmolVLM 256M0.256B
SMOLLM·8K CTX· CHAT· VISION
425
TOK/S · 4% VRAM
S
Gemma 3 270M0.27B
GEMMA·32K CTX· CHAT
403
TOK/S · 4% VRAM
S
Snowflake Arctic Embed M v2.00.305B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
357
TOK/S · 4% VRAM
S
bge-large-en-v1.5 335M0.335B
EMBEDDING·1K CTX· CHAT
325
TOK/S · 4% VRAM
S
mxbai-embed-large-v10.335B
EMBEDDING·1K CTX· EMBEDDING
325
TOK/S · 4% VRAM
S
Snowflake Arctic Embed L0.335B
EMBEDDING·1K CTX· EMBEDDING
325
TOK/S · 4% VRAM
S
LFM2 350M0.35B
LFM·125K CTX· CHAT· TOOL_USE
311
TOK/S · 4% VRAM
S
SmolLM2 360M0.36B
SMOLLM·8K CTX· CHAT
302
TOK/S · 4% VRAM
S
GPT-2 Medium 345M0.38B
GPT2·1K CTX· CHAT
286
TOK/S · 5% VRAM
S
Qwen 2.5 0.5B0.5B
QWEN·32K CTX· CHAT
218
TOK/S · 5% VRAM
S
SmolVLM 500M0.5B
SMOLLM·8K CTX· CHAT· VISION
218
TOK/S · 5% VRAM
S
Falcon-H1 0.5B0.52B
FALCON·128K CTX· CHAT
209
TOK/S · 5% VRAM
S
BGE-M30.568B
EMBEDDING·8K CTX· EMBEDDING
192
TOK/S · 5% VRAM
S
Snowflake Arctic Embed L v2.00.568B
EMBEDDING·8K CTX· EMBEDDING· MULTILINGUAL
192
TOK/S · 5% VRAM
S
Qwen3-Embedding 0.6B0.6B
EMBEDDING·32K CTX· EMBEDDING· MULTILINGUAL
181
TOK/S · 5% VRAM
S
Falcon-H1R Tiny 0.6B0.6B
FALCON·32K CTX· CHAT· REASONING
181
TOK/S · 5% VRAM
S
Falcon Perception 0.6B0.6B
FALCON·4K CTX· VISION
181
TOK/S · 5% VRAM
S
Qwen 1.5 0.5B0.62B
QWEN·32K CTX· CHAT
175
TOK/S · 5% VRAM
S
LFM2 700M0.74B
LFM·125K CTX· CHAT· TOOL_USE
147
TOK/S · 6% VRAM
S
Qwen3 0.6B0.75B
QWEN·32K CTX· CHAT· REASONING
145
TOK/S · 6% VRAM
S
GPT-2 Large 774M0.81B
GPT2·1K CTX· CHAT
134
TOK/S · 6% VRAM
S
Qwen 3.5 0.8B0.87B
QWEN·256K CTX· CHAT· CODING· MULTILINGUAL
125
TOK/S · 6% VRAM
S
InternVL3 1B0.94B
OTHER·32K CTX· CHAT· VISION
116
TOK/S · 7% VRAM
S
MiniCPM5 1B1.08B
MINICPM·128K CTX· CHAT· REASONING· MULTILINGUAL
101
TOK/S · 7% VRAM
S
TinyLlama 1.1B1.1B
LLAMA·2K CTX· CHAT
99
TOK/S · 7% VRAM
S
LFM2.5-1.2B-Thinking1.2B
LFM·122K CTX· CHAT· REASONING· TOOL_USE
91
TOK/S · 8% VRAM
S
Llama-3.2-1B1.2B
LLAMA·4K CTX· CHAT
91
TOK/S · 8% VRAM
S
LFM2 1.2B1.2B
LFM·125K CTX· CHAT· TOOL_USE· MULTILINGUAL
91
TOK/S · 8% VRAM
S
Zamba2 1.2B1.2B
OTHER·4K CTX· CHAT
91
TOK/S · 8% VRAM
S
EXAONE-4.0-1.2B1.3B
EXAONE·64K CTX· CHAT
84
TOK/S · 8% VRAM
S
OPT 1.3B1.3B
OPT·2K CTX· CHAT
84
TOK/S · 8% VRAM
S
MiniCPM-V 4.61.3B
OTHER·256K CTX· CHAT· VISION
84
TOK/S · 8% VRAM
S
DeepSeek Coder 1.3B1.35B
DEEPSEEK·16K CTX· CODING
81
TOK/S · 8% VRAM
S
Phi-1 1.3B1.42B
PHI·2K CTX· CODING
77
TOK/S · 8% VRAM
S
Phi-1.5 1.3B1.42B
PHI·2K CTX· CHAT· CODING
77
TOK/S · 8% VRAM
S
LFM2 8B A1B8.3BMoE
LFM·125K CTX· CHAT· TOOL_USE· MULTILINGUAL
73
TOK/S · 35% VRAM
▸ NEXT STEP

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Jetson AGX Xavier 16 GB.

WHAT THIS CARD IS WORTH

Jetson AGX Xavier 16 GB holds 262 of the models in our catalogue and is, in practice, a IQ4_XS card — the largest it takes is ERNIE 4.5 21B A3B at IQ4_XS.

TOKENS/SEC PER $100
8B at Q4_K_M, so cards compare like for like
VRAM PER $100
what memory costs on this card
THE BIGGEST IT TAKES
ERNIE 4.5 21B A3B21.95B · IQ4_XS13.9 GB54 tok/sEST
GPT-OSS 20B21B · Q4_K_S13.8 GB43 tok/sEST
Reka Flash 321B · Q4_K_S14.2 GB7 tok/sEST
Reka Flash 3.121B · Q4_K_S14.2 GB7 tok/sEST
InternLM2 20B19.8B · Q4_K_M14.3 GB8 tok/sEST
InternLM2.5 20B19.8B · Q4_K_M14.3 GB8 tok/sEST
Ling-lite 16.8B16.8B · Q5_K_M13.6 GB53 tok/sEST
DeepSeek V2 Lite 16B16B · Q5_K_M12.9 GB53 tok/sEST
AGAINST CARDS OF SIMILAR MEMORY
M1 Max (32GB)21 GB$1,4993.0 tok/s per $100
M2 Max (32GB)21 GB$1,7992.5 tok/s per $100
M2 Pro (32GB)21 GB$1,4991.9 tok/s per $100
M4 (32GB)21 GB$1,1991.7 tok/s per $100
Jetson AGX Xavier 16 GB16 GB

Neighbours in memory rather than in price: memory decides whether a card can do the job at all, so two cards of the same size at different prices is the comparison you are making. Ordered by memory, then bandwidth — not by the value column, which is worked out from bandwidth and price alone and therefore rewards a cheap card whatever its software stack does to that bandwidth in practice. Read it as one input, not as a ranking.

▸ DEVICE UNDER TEST

NVIDIA Jetson AGX Xavier 16 GB — 16 GB VRAM.

JETSON AGX XAVIER 16 GB SPEC
BRAND
NVIDIA
VRAM
16 GB LPDDR4X
BANDWIDTH
136 GB/s
FP16 COMPUTE
2.8 TFLOPS
FP32 COMPUTE
1.4 TFLOPS
CUDA CORES
512
TENSOR CORES
64
TDP
30 W
ARCHITECTURE
Volta
▸ AI CAPABILITY
262/ 449 models @ Q4

With 16 GB VRAM and 136 GB/s bandwidth, this GPU handles models up to 19.8B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~16 tok/s.

RENT IT — LIVE PRICES
Checking what the clouds are charging…
§ 01TOP MODELS FOR JETSON AGX XAVIER 16 GB
262 FIT · SHOWING 20
MODELSIZEVRAM Q4TOK/SAVG
InternLM2 20B19.8B12.6 GB545.1
InternLM2.5 20B19.8B12.6 GB550.9
Ling-lite 16.8B16.8B10.8 GB45
DeepSeek V2 Lite 16B16B10.3 GB4538.0
StarCoder2 15B15.96B10.2 GB726.5
DeepSeek-Coder-V2-Lite 15.7B15.7B10.1 GB4543.0
DeepSeek-VL2 Small 16B15.7B10.1 GB4543.1
StarCoder 15B15.5B10.0 GB721.0
InternVL3 14B15.12B9.7 GB738.1
Phi-4-reasoning-vision 15B15B9.7 GB742.8
DeepSeek R1 Distill Qwen 14B14.8B9.5 GB743.9
DeepCoder 14B14.8B9.5 GB738.7
Qwen2.5-Coder-14B14.8B9.5 GB741.3
Qwen2.5-14B14.8B9.5 GB741.3
Qwen3 14B14.8B9.5 GB745.7
phi-4 14B14.66B9.4 GB733.7
Phi-4-reasoning 14B14.66B9.4 GB733.7
Phi-4-reasoning-plus 14B14.66B9.4 GB775.5
Ministral 3 14B14B9.0 GB825.9
Phi-3-medium-14b14B9.0 GB833.7