Models ready for routing.
OPS-09
Models & Compute
Model status grid, selected model config snapshot, keep-alive, fallback route and recent errors aligned to the v3 wireframe.
Tracetrc_models_y3j0e5Generated11:47:23Models4
Models behind saturation or budget guardrails.
Average latency across all visible models.
Combined queue depth on the compute layer.
Model status grid
Name, host, queue and latency.
local-llama-70b
Keep-alive, fallback route and recent errors.
Host mac-studio-mainParallel limit 8Keep-alive 15m
Fallback route
gpt-5.2-mini
Config snapshot
- context 64k
- quant q6_k
- semantic-recall=true
Recent errors
- none in last 6h
Routing matrix
Primary-to-fallback mapping required by the models contract.
| Model | Host | Fallback | Queue depth |
|---|---|---|---|
| local-llama-70b | mac-studio-main | gpt-5.2-mini | 6 |
| gpt-5.2 | premium-remote | gpt-5.2-mini | 2 |
| claude-sonnet | premium-remote | local-llama-70b | 4 |
| local-mistral | vps-prod-2 | gpt-5.2-mini | 2 |
Rollback action modal
Config rollback readiness and guarded change notes.
Selected config snapshot
- context 64k
- quant q6_k
- semantic-recall=true
Recent errors
- none in last 6h