What can I run on Jetson AGX Orin 64GB?
64 GB unified memoryAmpere (2048 CUDA cores, 64 Tensor Cores)275 TOPS15W - 60W (nvpmodel modes 15W / 30W / 50W / MAXN)View matrix for this target →
AGX Orin 64GB module / developer kit. Distinct from AGX Orin 32GB (fewer cores, 40W max mode) and AGX Orin Industrial. Of the Orin modules in NVIDIA's power-mode tables, only the AGX Orin 64GB has a 50W nvpmodel mode. Record power mode with every run. 64 GB is unified memory shared with the OS.
Policy by policy
| Policy | Params | Inference memory (sourced, with precision) | On Jetson AGX Orin 64GB | Best measured |
|---|---|---|---|---|
| π0.5 VLA | — | ≥8 GB | measured | 3.69 Hz · llama.cpp · fp16 · 2 cam |
| ACT Imitation | 80M | 211 MB measured fp32 | no data | |
| Diffusion Policy Imitation | — | 1.12 GB measured fp32 | no data | |
| EO-1 VLA | 3B | ≥6.5 GB | no data | |
| Evo-1 VLA | 770M | — | no data | |
| Fast-WAM World model | 6B | — | no data | |
| GR00T N1.7 VLA | 3B | ≥16 GB | no data | |
| HIL-SERL RL | — | — | no data | |
| LingBot-VA World model | 5.3B | ≥18 GB | no data | |
| MolmoAct2 VLA | — | 12.1 GB measured bf16 | no data | |
| Multitask DiT Imitation | 450M | — | no data | |
| OpenVLA VLA | 7.5B | 15 GB measured bf16 | no data | |
| OpenVLA-OFT VLA | — | ≥16 GB | no data | |
| SmolVLA VLA | 450M | 1.75 GB measured fp32 | no data | |
| TD-MPC RL | — | — | no data | |
| VLA-JEPA VLA | — | — | no data | |
| VQ-BeT Imitation | — | — | no data | |
| WALL-OSS VLA | — | — | no data | |
| X-VLA VLA | 900M | — | no data | |
| π0 VLA | 3.3B | 13.32 GB measured fp32 | no data | |
| π0-FAST VLA | — | ≥8 GB | no data |
“Exceeds memory” compares a policy's upstream-stated minimum inference memory (never a measured peak, which is precision-specific) with this target's unified (shared CPU/GPU) memory, so the real budget is lower. It is arithmetic, not observation, and is labelled inferred. Quantised or split runtimes can change the answer; a measured run always wins.
All measured runs on Jetson AGX Orin 64GB
| Policy | Runtime | Precision | Cams | p50 / p95 ms | Inference | Platform | Source |
|---|---|---|---|---|---|---|---|
| π0.5 | llama.cpp | fp16 | 2 @224x224 | 271.1 / — | 3.69 Hz | power_mode 50W (paper label "Orin-50W")conditionsboard: Jetson Orin, 'Orin-50W' power mode (AGX Orin 64GB inferred, see ev-orin-board-inferred) Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). Later configuration from the EagleVLA-Edge README (time cache + cuBLASLt GEMM tuning, PR #6); not in the paper. Naive PyTorch π0.5 on the same board per the paper: 1420.8 ms (0.70 Hz) (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
| π0.5 | llama.cpp | fp16 | 2 @224x224 | 374.5 / — | 2.67 Hz | power_mode 50W (paper label "Orin-50W")conditionsboard: Jetson Orin, 'Orin-50W' power mode (AGX Orin 64GB inferred, see ev-orin-board-inferred) Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). Paper's 'control frequency' (6.93 Hz) is 1/reaction time with confidence-based scheduling that skips the VLM; it is not executed actions/s and is not recorded as control_rate_hz. Naive PyTorch π0.5 on the same board per the paper: 1420.8 ms (0.70 Hz) (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
Sources
- T1 · upstream documentedvendor-docsverified 2026-10-24 (1d ago)
NVIDIA Jetson Orin spec table AI Performance row. Column order on the page: AGX Orin Developer Kit / AGX Orin 64GB (275), AGX Orin Industrial (248), AGX Orin 32GB (241), Orin NX 16GB (157), Orin NX 8GB (117), Orin Nano Super Developer Kit (67), Orin Nano 8GB (67), Orin Nano 4GB (34).
“AI Performance 275 TOPS 248 TOPs 241 TOPS 157 TOPS 117 TOPS 67 TOPS 67 TOPS 34 TOPS”
https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/ - T1 · upstream documentedvendor-docsverified 2026-10-24 (1d ago)
Jetson Orin spec table Memory row starts with the AGX Orin 64GB column: 64GB 256-bit LPDDR5, 204.8 GB/s.
“Memory 64GB 256-bit LPDDR5 204.8GB/s”
https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/ - T1 · upstream documentedvendor-docsverified 2026-10-24 (1d ago)
Jetson Orin spec table Power row; first column (AGX Orin Developer Kit / 64GB) is 15W - 60W.
“Power 15W - 60W 15W-75W 15W - 60W 10W - 15W - 25W - 40W”
https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/ - T1 · upstream documentedofficial-docsverified 2026-10-24 (1d ago)jetson_linux 36.4.3
Jetson Linux r36.4.3 power-mode table for AGX Orin 64GB: MAXN, 15W, 30W (default), 50W.
“for Jetson AGX Orin 64GB Property Mode MAXN 15W 30W* 50W Power budget n/a 15W 30W 50W”
https://docs.nvidia.com/jetson/archives/r36.4.3/DeveloperGuide/SD/PlatformPowerAndPerformance/JetsonOrinNanoSeriesJetsonOrinNxSeriesAndJetsonAgxOrinSeries.html - T1 · upstream documentedofficial-docsverified 2026-10-24 (1d ago)jetson_linux 36.4.3
Same guide: AGX Orin 32GB modes are MAXN, 15W, 30W, 40W (no 50W mode).
“NVP Model Clock Configuration for Jetson AGX Orin 32GB Property Mode MAXN 15W 30W* 40W Power budget n/a 15W 30W 40W”
https://docs.nvidia.com/jetson/archives/r36.4.3/DeveloperGuide/SD/PlatformPowerAndPerformance/JetsonOrinNanoSeriesJetsonOrinNxSeriesAndJetsonAgxOrinSeries.html