AGORA
Nodes
106/ 555
Contributors
0/ 98
Peak
118
Loss
2.432
TPS
23 kTPS
Protocol overhead
90%
Pluralis
16%
Join queue
0queued · ETA 4 min
Showing single run
Pluralis-8BLive
Pluralis-8BLiveModelPluralis-8BDatasetFineWeb-Edu (1.3T)
Join queue
0queued
Join queue
Pipeline0 Stages · 0 Online · 0/98 Contributors
No swarm data
Range
Smooth0.00
X
Training loss
Cross-entropy
Throughput
Tokens / sec
Throughput per TFLOP
Tok/s divided by active dense BF16 TFLOPs · higher is better
0.65
Baselines: optimal central setting 6.74 and Megatron-LM at 200 Mbps 0.37 tok/s/TFLOP. See Analysis tab for full breakdown + sources.
Total swarm TFLOPS
Active dense BF16 compute · Pluralis-operated vs Contributor split
36.37kPluralis 16%Contributor 84%
Mean MFU
Model FLOP utilisation · achieved ÷ peak · higher is better (Chinchilla 6N, 8B, denom: BF16)
3.1%
Evaluations
No data yet for Pluralis-8B
Benchmark results will appear here as the training run progresses.