🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the Jetson Thor T2000 run Surya?
Yes — here's how fast and up to what context, with real measured numbers.
Surya on the Jetson Thor T2000
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 186k | 136.6-162.8 t/s | ~10-11.9 (slow) |
| Q8 | yes | 191k | 273.3-325.7 t/s | ~19.9-23.8 (slow) |
| Q4 | yes | 194k | 487.9-581.5 t/s | ~35.6-42.4 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the Jetson Thor T2000 run Surya?▶
Yes. At Q4 it generates ~487.9-581.5 tok/s and fits up to 194k context; beyond that it spills to SSD swap and slows to ~35.6-42.4 tok/s.
How much context fits?▶
At Q4, up to about 194k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.