Inference
Round--:--

Your GPU runs the model. The fees pay for the tokens.

Your browser loads a real language model onto your graphics card and answers the pool's prompts. Every 10 minutes the $INFERENCE creator fees are split by the tokens each machine generated.

Your GPU
Checking…Checking your graphics card…
Modelf16 build

Runs on your graphics card through WebGPU. Keep the tab open; close it to stop.

Your speed
0tok/s
Model
–
Backend
none
Last answer
–
Peak
–
Session
–
Tokens this session
0
Average speed
–
Longest answer
–
VRAM for this model
945 MB
Not counted
0
Your tokens this round00 in the pool
Your share of the round0%Hold 100k to be paid
Answers this session00 tokens this session
Round ends in--:--0 machines this round

Your model

The prompt your GPU is answering now, written token by token.

idle
prompt → model → tokens

Press Start. Your model answers here, token by token.

Network answers

Every answer the pool accepted, newest first.

0 this round

No answers yet. Start your model and yours is the first one here.

The pool right now

Live from the board, the AI compute market and Solana.

Pool speed0 tok/s0 machines active
Tokens this round00 answers
Fastest GPU–none this round
AI compute–loading
AI compute–loading
AI compute–loading
Inference$INFERENCE