1 build of Sarvam-M measured on real machines — different packagers, precisions and engines, each a different file with its own numbers. The best each engine proved is first; every build has its own page with the machines and signatures behind it. 24B parameters, as the vendor states.
trained 32K · proved 8K context proved by bartowski Q4_K_M
15.9 tok/s peak, by bartowski Q4_K_M
proved by at least one build: Sustained output · Extended thinking · Reasons about code · Synthesises sources
| build | publisher | precision | engine | whose work | context proved | best tok/s | machines |
|---|---|---|---|---|---|---|---|
| hf.co/bartowski/sarvamai_sarvam-m-GGUF:Q4_K_M | bartowski | Q4_K_M | ollama | quantizer | 8,192 | 15.9 | 2 |