
| Jetson AGX Thor T4000 | ||
| AI Performance | 1200 TFLOPS (FP4—Sparse) | |
| GPU | 1536-core NVIDIA Blackwell architecture GPU with fifth-gen Tensor Cores Multi-Instance GPU (MIG) with six TPCs | |
| CPU | 12-core Arm® Neoverse®-V3AE 64-bit CPU 1 MB L2 cache per core 16 MB shared system L3 cache | |
| Memory | 64 GB 256-bit LPDDR5X 273 GB/s | |
| Video Encode | 1x NVENCODE 3x 4Kp60 (H.265) | 6x 4Kp30 (H.265) | 12x 1080p60 (H.265) | 25x 1080p30 (H.265) | 24x 1080p30 (H.264) | 3x 4Kp60 (H.264) | |
| Video Decode | 1x NVDECODE 2x 8Kp30 (H.265) | 5x 4Kp60 (H.265) | 11x 4Kp30 (H.265) | 23x 1080p60 (H.265) | 46x 1080p30 (H.265) | 41x 1080p30 (H.264) | 2x 4Kp60 (H.264) | |
| Power | 40W-70W | |
| HSCODE | 8471504090 |
| USHSCODE | 8517180050 |
| UPC | |
| EUHSCODE | 8471707000 |
| COO | CHINA |
| NVIDIA AGX Thor T4000 Module | x1 |
NVIDIA® Jetson Thor™ T4000 modules deliver up to 120 FP4 TFLOPS of AI compute and 64 GB of memory — 7.5× the performance and 3.5× the energy efficiency of NVIDIA AGX Orin™ — at 40–70 W. They support a wide range of popular open-source models through Jetson AI Lab.
NVIDIA® Jetson AGX Thor™ T4000 is not available for purchase separately, please select it from the bundle.