{"version":"2.1.314","anchor":"effort-level-400-handling-gated-by-tengu-eager-quilt","canonical_anchor":"effort-level-400-handling-gated-by-tengu-eager-quilt","heading":"Claude Code can step down the effort level when a model rejects it","tier":"notice","area":"Effort","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.314\/e\/effort-level-400-handling-gated-by-tengu-eager-quilt","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.314","markdown":"### Claude Code can step down the effort level when a model rejects it\n\nWhen a model rejects an effort level, Claude Code can retry one level lower and tell you once, unless a remote switch turns this off\n\n**Unclear.** It is not clear which requests use this fallback.\n\n**What**\n\nThe effort level is the setting that controls how much work the model puts into a reply. When the model service rejects a request with an \"Invalid effort level\" error, Claude Code can now treat it as an error it can recover from:\n\n- It lowers the effort level by one step and tries again.\n\n- It remembers the lower level for that model.\n\n- It tells you once that the level was lowered.\n\nThis recovery stays on unless a remote switch turns it off. That switch is off by default.\n\n**Why**\n\nIf a model does not accept the effort level you picked, the request can still go through at the nearest level below instead of failing.\n\n- Flag `tengu_eager_quilt`: Not enough to say (read for one account on one subscription tier against v2.1.314; this account: no value returned, anonymous baseline: no value returned, compiled default: off) These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.\n- Area: Effort\n- Tier: You'll notice\n- Useful: 3\/5\n- Signal: 3\/5\n- Scope: individual\n- Heads-up: no"}