{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiggl57wr6mxstrt4t2zrhpmrkx6mvag4ni7avdncnopywooipymk4",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mofp5jd4wgv2"
  },
  "path": "/t/selected-model-is-at-capacity-please-try-a-different-model-june-16th-2026/1383863?page=2#post_38",
  "publishedAt": "2026-06-16T11:09:48.000Z",
  "site": "https://community.openai.com",
  "textContent": "I think the goblins may need a ledger.\n\nI’m seeing repeated failures from **Codex via CLI** when using **GPT-5.5 with xhigh reasoning effort**.\n\nThe error is:\n\n> Selected model is at capacity. Please try a different model.\n\nEnvironment:\n\nSurface: Codex CLI\nPlatform: Linux\nPlan: ChatGPT Pro\nModel: GPT-5.5\nReasoning effort: xhigh\nDate observed: 16 June 2026\nFrequency: repeated / intermittent, but common enough to interrupt normal use\n\nThis seems consistent with the current OpenAI Status incident for Codex, rather than a local connectivity or authentication problem.\n\nA few questions that might help make this clearer for CLI users:\n\n  1. Is this capacity error currently expected specifically for **GPT-5.5 xhigh** , or is it affecting Codex more broadly?\n\n  2. When the CLI says “try a different model”, does that mean:\n\n     * switch model,\n\n     * reduce reasoning effort from `xhigh` to `high` or `medium`,\n\n     * or either/both?\n\n  3. Is there currently a CLI-visible way to distinguish between:\n\n     * model saturation,\n\n     * account quota exhaustion,\n\n     * Codex service degradation,\n\n     * routing issues,\n\n     * or local/client-side configuration problems?\n\n  4. Could Codex CLI expose a clearer diagnostic code or category here? The current message is readable, but it does not give much guidance on whether the user should retry, change effort level, switch model, check quota, or wait for incident mitigation.\n\n  5. If `xhigh` is capacity constrained, could the CLI offer an automatic fallback prompt, such as:\n\n\n\n\n> xhigh unavailable — retry with high?\n\nThat would be more helpful than failing the session outright.\n\nI guess either the goblins are striking, or the hamster on the server wheel needs reinforcements.\n\nFor CLI users, especially those using Codex for governed or review-heavy engineering work, a generic “selected model is at capacity” message creates ambiguity. It would be helpful to know whether this is transient service capacity, `xhigh`-specific saturation, or something users are expected to manage through configuration.\n\nHappy to call a performance meeting with the goblins if required.",
  "title": "Selected Model is at capacity. Please try a different model. June 16th, 2026"
}