P
Manufacturer
Poolside
Laguna-Coding-Modelle von Poolside.
Best token generation
1.765,0 tok/s
Laguna-XS-2.1
Best prefill
39.157,1 tok/s
Laguna-XS-2.1
∅ Token generation
361,4 tok/s
Average of 271 runs
∅ Prefill
4.683,0 tok/s
Average of 271 runs
Best TTFT
212 ms
Laguna-XS-2.1
Best harness rate
0%
no harness run yet
🏆
Laguna-XS-2.1 · fastest model at 1.765,0 tok/s token generation
Best prefill of this manufacturer: 39.157,1 tok/s (Laguna-XS-2.1)
Best prefill of this manufacturer: 39.157,1 tok/s (Laguna-XS-2.1)
Top models by token generation
Best measured generation throughput per model (tok/s), published performance benchmarks.
Models
All benchmarks →6 assigned
Laguna-M.1225.8B77 runs
39,7tok/s Ø
Min 0,3Max 337,6
Laguna-M.1-GGUF-Q4_K_M225.8B0 runs
No performance data yet
Laguna-S-2.1117.6B44 runs
222,4tok/s Ø
Min 4,8Max 835,1
Laguna-S-2.1-INT4117.6B21 runs
424,3tok/s Ø
Min 9,1Max 781,4
Laguna-XS-2.133.4B127 runs
599,5tok/s Ø
Min 29,7Max 1.765,0
Laguna-XS-2.1-INT433.4B2 runs
26,9tok/s Ø
Min 26,9Max 26,9

