VLM Inference Benchmark Explorer
Alpamayo2-Super on Jetson AGX Thor — BF16, vLLM vision-language inference benchmark
Open this configuration in the VLM Inference Benchmark Explorer
| Model | Alpamayo2-Super |
|---|---|
| Parameters | 34.3B |
| Device | Jetson AGX Thor |
| Quantization | BF16 |
| Image size | 720p |
| Images per camera | 1 |
| Max Cameras (TTFT ≤ 3000 ms) | 16 |
| Engine | vLLM |
| Cameras | TTFT (ms) | TPS (tok/s) | Target |
|---|---|---|---|
| 1 | 780 | 4.6 | ≤3000 ms |
| 2 | 1061 | 4.8 | ≤3000 ms |
| 4 | 1164 | 4.6 | ≤3000 ms |
| 8 | 1534 | 4.4 | ≤3000 ms |
| 16 | 2154 | 3.9 | ≤3000 ms |
| 32 | 3254 | 2.9 | >3000 ms |
| 64 | 5423 | 1.7 | >3000 ms |
| 128 | 14152 | 1.5 | >3000 ms |
| 160 | 18384 | 1.3 | >3000 ms |
| 176 | 20711 | 1.4 | >3000 ms |
| 192 | — | — | Failed |
| Images per request | TTFT (ms) | TPS (tok/s) | Target |
|---|---|---|---|
| 1 | 780 | 4.6 | ≤3000 ms |
| 2 | 833 | 4.6 | ≤3000 ms |
| 4 | 1064 | 4.8 | ≤3000 ms |
| 8 | 1711 | 5.2 | ≤3000 ms |
| 16 | 3364 | 5.7 | >3000 ms |
| 32 | 6442 | 217.8 | >3000 ms |
| 64 | 14952 | 212.3 | >3000 ms |
| 128 | 35365 | 6.7 | >3000 ms |
| 256 | 135664 | 3.7 | >3000 ms |
| 512 | — | — | Failed |
Alpamayo2-Super (34.3B parameters), served in BF16 format with vLLM on Jetson AGX Thor, starts answering a single camera that sends one 720p image after 780 ms (time to first token, TTFT), then writes 4.6 tok/s.
With 256 images per camera at 720p, the answer starts after 135,664 ms. With one 2K image, after 838 ms.
At the default target of a mean TTFT within 3,000 ms (720p, one image per camera), it keeps up with 16 cameras at once.
Alpamayo2-Super (34.3B parameters), served in BF16 format with vLLM on Jetson AGX Thor, starts answering a single camera that sends one 720p image after 780 ms (time to first token, TTFT), then writes 4.6 tok/s.
With 256 images per camera at 720p, the answer starts after 135,664 ms. With one 2K image, after 838 ms.
At the default target of a mean TTFT within 3,000 ms (720p, one image per camera), it keeps up with 16 cameras at once.
Alpamayo2-Super (a 32B Qwen3-VL backbone and a 2.3B diffusion expert) writes a short chain of causation from six cameras (4 frames each) and the vehicle's recent motion, then plans a trajectory. The table above measures only its vision-language part, served with vLLM in BF16. To "Describe the scene." that part mostly answers in a few tokens (2 to 34 on average per cell) that no scene description was provided, although it does describe the image when asked what it sees ("The scene shows a road intersection"); its tok/s figures therefore swing more than other rows'. This test runs the whole model with the official inference path of its repository (NVlabs/alpamayo2, inference_smoke.py: BF16, 10 diffusion steps, one trajectory, top-p 0.98, temperature 0.6) on the repository's 5 validation clips, 3 runs each, Jetson AGX Thor at MAXN.
In the table above 2 requests came back with an empty answer; as in the camera limit, an empty answer is not counted as a failure, so it does not end a series. There is no TensorRT recipe for this model; attention runs through PyTorch SDPA, because flash-attn has no build for this board.
| Step | Mean | p50 | p90 |
|---|---|---|---|
| Whole decision | 13.3 | 12.6 | 17.5 |
| Input preparation | 1.0 | 1.0 | 1.0 |
| Model (chain of causation and trajectory) | 12.3 | 11.7 | 16.5 |
Each decision reads 24 images (4,580 prompt tokens) and writes about 18 tokens of chain of causation (10 to 32); the longer the chain, the longer the decision. The model takes 69 GiB of memory. Alpamayo2-Edge, with its TensorRT recipe on the same board, takes 331 ms per decision.