{"version":"2.1.323","anchor":"max-output-tokens-env-style-int-bound-raised-from-64000-to-3","canonical_anchor":"max-output-tokens-env-style-int-bound-raised-from-64000-to-3","heading":"A whole-number setting now accepts values up to 300000","tier":"notice","area":"Settings","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.323\/e\/max-output-tokens-env-style-int-bound-raised-from-64000-to-3","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.323","markdown":"### A whole-number setting now accepts values up to 300000\n\nA setting that takes a whole number from 32000 upward now accepts values up to 300000 instead of stopping at 64000\n\n**Unclear.** Which setting this limit applies to is not identified, so it is not confirmed that it is the maximum output tokens setting.\n\n**What**\n\nA setting that accepts only digits, with a minimum of 32000, has a higher upper limit. Its largest allowed value went from 64000 to 300000.\n\n**Why**\n\nIf this is the limit on how many tokens Claude Code lets the model write in one reply, you can now set it far higher than before. Tokens are the small chunks of text the model reads and writes.\n\n- Area: Settings\n- Tier: You'll notice\n- Useful: 2\/5\n- Signal: 2\/5\n- Scope: individual\n- Heads-up: no"}