Models
Compare capabilities, context limits and credit rates.
Service running
Error in the last hour; usable
Model or probe not working
Probe failed; can try
Available model directory
ModelInputContextMax outputInput rateCache rateOutput rateLive status
Loading models...
Latency overview
Gateway + upstream combined · successful requests only
Latest request: Not recorded
Model comparison
Each model’s own latest 200 requests · time windows may differ
| Model | Average latency | Measured |
|---|---|---|
| Loading latency… | ||
Up to 10 public models selected by request volume in the global latest 200. Measured = successful requests with latency / all sampled requests. Full response duration, not time to first token.