Models for reviews
Keep model choices understandable.
The model catalog retains current and historical choices. Pricing and benchmark data stay with their dated sources in Token Economy.
Model catalog
Current review routes, with history intact.
Quality Studio uses the models offered by the installed agent CLIs. The review toolchain default is gpt-5.6-luna; it replaced gpt-5.4-mini after that model was withdrawn on 2026-09-09. Current models are fully visible. Historical models remain as one collapsed row per model.
| Model | CLI | Status | Quality Studio role |
|---|---|---|---|
claude-sonnet-5 | claude | current | available route |
claude-opus-5 | claude | current | available route |
claude-fable-5-1 | claude | current | available route |
gpt-5.6-sol | codex | current | available route |
gpt-5.6-terra | codex | current | available route |
gpt-5.6-luna | codex | current | default review route |
gpt-6-astra | codex | current | available route |
| |||
| |||
| |||
| |||
| |||
| |||
| |||
| |||
| |||
| |||
| |||
| |||
| |||
Catalog and review default audited 2026-09-12 against the installed CLI inventory and operator routing policy. Price and score evidence is not duplicated here; use the dated Token Economy benchmark matrix (source snapshot 2026-09-11).