⌬
Cortex
Demo
Gallery
Personas
Specs
Seratonin
RTX 5090
Big Apple
M4 Max
OpenRouter
cloud failover
GPU: idle
Connecting…
⚡ Local · RTX 5090
Ant
Post
L Lat
R Lat
Sup
Inf
Both
LH
RH
+z
−z
▶
t = 0.0 s
⌬
Submit a media file to begin.
TRIBE v2 will predict cortical BOLD responses. Gemma will narrate.
⚡
Local Pipeline
RTX 5090 · Gemma 4 E4B · TRIBE v2 local
575 W
GPU TDP estimate
$0.15 / kWh
electricity rate
—
avg ¢ per run
—
total ¢ this session
Formula:
W × seconds ÷ 3600 ÷ 1000 × rate × 100
Scan
File
TRIBE (s)
Gemma (s)
¢
☁
Cloud Pipeline
Vertex AI · Gemma 4 27B · TRIBE v2 cloud GPU
A100 GPU
cloud inference
$3.67 / hr
A100 on-demand rate
~3.1 ¢
avg ¢ per run (est.)
—
projected ¢ this session
TRIBE inference (cloud A100, ~30s)
≈ 3.06 ¢
Gemma 4 27B input (~600 tok)
≈ 0.021 ¢
Gemma 4 27B output (~400 tok)
≈ 0.042 ¢
Total per run
≈ 3.12 ¢
Scan
File
Local ¢
Cloud ¢ (est.)
Savings