{"version":"2.1.199","anchor":"model-catalog-max-output-tokens-limits-and-advisor-rank","canonical_anchor":"model-catalog-max-output-tokens-limits-and-advisor-rank","heading":"Model Catalog: max_output_tokens Limits and advisor_rank","tier":null,"area":null,"scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.199\/e\/model-catalog-max-output-tokens-limits-and-advisor-rank","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.199","markdown":"### Model Catalog: max_output_tokens Limits and advisor_rank\n\nModels in the internal catalog now carry explicit output token limits and an advisor ranking.\n\n**Details**\n\n- `max_output_tokens.default` and `max_output_tokens.upper` are now populated for each model (e.g., Haiku 3.5: 8192\/8192; Haiku 4.5: 32000\/64000)\n- `advisor_rank` field added to models eligible for the advisor tool; models without an advisor rank emit a clearer error message pointing to `CLAUDE_CODE_ENABLE_EXPERIMENTAL_ADVISOR_TOOL=1`\n\n**Evidence**\n\n(search for `\"max_output_tokens\"`, `\"advisor_rank\"`, `\"has no advisor rank in the model catalog\"`)"}