Unintended Fast-Mode Billing on Codex Subagents
What version of Codex CLI is running?
0.149.0
What subscription do you have?
Pro 5X
Which model were you using?
gpt-5.6-sol, gpt-5.6-terra
What platform is your computer?
Linux 7.0.0-29-generic x86_64 unknown (Ubuntu 26.04)
What terminal emulator and version are you using (if applicable)?
WezTerm
Codex doctor report
{
"schemaVersion": 1,
"generatedAt": "1787301659s since unix epoch",
"overallStatus": "ok",
"codexVersion": "0.149.0",
"checks": {
"app_server.status": {
"id": "app_server.status",
"category": "app-server",
"status": "ok",
"summary": "background server is not running",
"details": {
"control socket": "/home/gweeteve/.codex/app-server-control/app-server-control.sock",
"daemon state dir": "/home/gweeteve/.codex/app-server-daemon",
"mode": "ephemeral",
"pid file": "/home/gweeteve/.codex/app-server-daemon/app-server.pid (missing)",
"settings": "/home/gweeteve/.codex/app-server-daemon/settings.json (missing)",
"status": "not running",
"update-loop pid file": "/home/gweeteve/.codex/app-server-daemon/app-server-updater.pid (missing)"
},
"remediation": null,
"durationMs": 0
},
"auth.credentials": {
"id": "auth.credentials",
"category": "auth",
"status": "ok",
"summary": "auth is configured",
"details": {
"auth file": "/home/gweeteve/.codex/auth.json",
"auth storage mode": "File",
"stored API key": "false",
"stored ChatGPT tokens": "true",
"stored agent identity": "false",
"stored auth mode": "chatgpt"
},
"remediation": null,
"durationMs": 0
},
"config.load": {
"id": "config.load",
"category": "config",
"status": "ok",
"summary": "config loaded",
"details": {
"CODEX_HOME": "/home/gweeteve/.codex",
"config.toml": "/home/gweeteve/.codex/config.toml",
"config.toml parse": "ok",
"cwd": "/home/gweeteve",
"enabled feature flags": "shell_tool, view_image, unified_exec, shell_snapshot, code_mode_host, terminal_resize_reflow, sqlite, memories, hooks, enable_request_compression, unbounded_connection_retries, multi_agent, apps, tool_search_always_defer_mcp_tools, tool_suggest, plugins, in_app_browser, in_app_chat, in_app_dictation, in_app_updates, browser_use, browser_use_full_cdp_access, browser_use_external, computer_use, remote_plugin, plugin_sharing, image_generation, resize_all_images, item_ids, skill_mcp_dependency_install, skill_search, mentions_v2, steer, guardian_approval, goals, collaboration_modes, tool_call_mcp_elicitation, auth_elicitation, personality, fast_mode, tui_app_server, remote_compaction_v2, workspace_dependencies",
"feature flag overrides": "memories=true",
"feature flags enabled": "43",
"log dir": "/home/gweeteve/.codex/log",
"mcp servers": "0",
"model": "gpt-5.6-sol",
"model provider": "openai",
"sqlite home": "/home/gweeteve/.codex"
},
"remediation": null,
"durationMs": 0
},
"git.environment": {
"id": "git.environment",
"category": "git",
"status": "ok",
"summary": "git version 2.53.0",
"details": {
"PATH git #1": "/usr/bin/git",
"PATH git #2": "/bin/git",
"PATH git entries": "2",
"git build options": "git version 2.53.0; cpu: x86_64; no commit associated with this build; sizeof-long: 8; sizeof-size_t: 8; shell-path: /bin/sh; rust: disabled; gettext: enabled; libcurl: 8.18.0; zlib: 1.3.1; SHA-1: SHA1_DC; SHA-256: SHA256_BLK; default-ref-format: files; default-hash: sha1",
"git exec path": "/usr/lib/git-core",
"git version": "git version 2.53.0",
"repo detected": "false",
"selected git": "/usr/bin/git"
},
"remediation": null,
"durationMs": 57
},
"installation": {
"id": "installation",
"category": "install",
"status": "ok",
"summary": "installation looks consistent",
"details": {
"PATH codex #1": "/home/gweeteve/.nvm/versions/node/v24.15.0/bin/codex",
"current executable": "/home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/bin/codex",
"install context": "npm (package /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl, bin /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/bin, resources /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/codex-resources, path /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/codex-path)",
"managed by bun": "false",
"managed by npm": "true",
"managed by pnpm": "false",
"managed package root": "/home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex",
"npm update target": "/home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex"
},
"remediation": null,
"durationMs": 115
},
"mcp.config": {
"id": "mcp.config",
"category": "mcp",
"status": "ok",
"summary": "no MCP servers configured",
"details": {},
"remediation": null,
"durationMs": 0
},
"network.env": {
"id": "network.env",
"category": "network",
"status": "ok",
"summary": "network-related environment looks readable",
"details": {
"managed proxy": "not configured",
"proxy env vars": "none",
"respect system proxy": "disabled"
},
"remediation": null,
"durationMs": 0
},
"network.provider_reachability": {
"id": "network.provider_reachability",
"category": "reachability",
"status": "ok",
"summary": "active provider endpoints are reachable over HTTP",
"details": {
"ChatGPT inference URL": "https://chatgpt.com/backend-api/<redacted> reachable (HTTP 405)",
"reachability mode": "ChatGPT auth"
},
"remediation": null,
"durationMs": 132
},
"network.websocket_reachability": {
"id": "network.websocket_reachability",
"category": "websocket",
"status": "ok",
"summary": "Responses WebSocket handshake succeeded",
"details": {
"DNS": "2 IPv4, 2 IPv6, first IPv6",
"auth mode": "chatgpt",
"connect timeout": "15000 ms",
"endpoint": "wss://chatgpt.com/backend-api/<redacted>",
"handshake result": "HTTP 101 Switching Protocols",
"model provider": "openai",
"provider name": "OpenAI",
"proxy env vars": "none",
"reasoning header": "false",
"server model present": "false",
"supports websockets": "true",
"wire API": "responses"
},
"remediation": null,
"durationMs": 573
},
"runtime.provenance": {
"id": "runtime.provenance",
"category": "runtime",
"status": "ok",
"summary": "running npm on linux-x86_64",
"details": {
"commit": "unknown",
"current executable": "/home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/bin/codex",
"install method": "npm (package /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl, bin /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/bin, resources /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/codex-resources, path /home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/codex-path)",
"platform": "linux-x86_64",
"version": "0.149.0"
},
"remediation": null,
"durationMs": 0
},
"runtime.search": {
"id": "runtime.search",
"category": "search",
"status": "ok",
"summary": "search is OK (bundled)",
"details": {
"search command": "/home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex/node_modules/@openai/codex-linux-x64/vendor/x86_64-unknown-linux-musl/codex-path/rg",
"search command readiness": "file exists",
"search provider": "bundled"
},
"remediation": null,
"durationMs": 0
},
"sandbox.helpers": {
"id": "sandbox.helpers",
"category": "sandbox",
"status": "ok",
"summary": "sandbox configuration is readable",
"details": {
"approval policy": "OnRequest",
"codex-linux-sandbox helper": "/home/gweeteve/.codex/tmp/arg0/codex-arg07k8FYn/codex-linux-sandbox",
"execve wrapper helper": "/home/gweeteve/.codex/tmp/arg0/codex-arg07k8FYn/codex-execve-wrapper",
"filesystem sandbox": "restricted",
"network sandbox": "restricted"
},
"remediation": null,
"durationMs": 0
},
"security.endpoint": {
"id": "security.endpoint",
"category": "security",
"status": "ok",
"summary": "endpoint protection is not inspected on this platform",
"details": {
"endpoint products": "not inspected on this platform"
},
"remediation": null,
"durationMs": 0
},
"state.paths": {
"id": "state.paths",
"category": "state",
"status": "ok",
"summary": "state paths and databases are inspectable",
"details": {
"CODEX_HOME": "/home/gweeteve/.codex (dir)",
"active rollout files": "649 files, 1095493279 total bytes, 1687971 average bytes",
"archived rollout files": "0 files, 0 total bytes, 0 average bytes",
"goals DB": "/home/gweeteve/.codex/goals_1.sqlite (file)",
"goals DB integrity": "ok",
"log DB": "/home/gweeteve/.codex/logs_2.sqlite (file)",
"log DB integrity": "ok",
"log dir": "/home/gweeteve/.codex/log (dir)",
"memories DB": "/home/gweeteve/.codex/memories_1.sqlite (file)",
"memories DB integrity": "ok",
"queue DB": "/home/gweeteve/.codex/queue_1.sqlite (file)",
"queue DB integrity": "ok",
"sqlite home": "/home/gweeteve/.codex (dir)",
"state DB": "/home/gweeteve/.codex/state_5.sqlite (file)",
"state DB integrity": "ok",
"thread history DB": "/home/gweeteve/.codex/thread_history_1.sqlite (file)",
"thread history DB integrity": "ok"
},
"remediation": null,
"durationMs": 398
},
"state.rollout_db_parity": {
"id": "state.rollout_db_parity",
"category": "threads",
"status": "ok",
"summary": "rollout files and state DB thread inventory agree",
"details": {
"default model provider": "openai",
"rollout DB active files": "649",
"rollout DB active rows": "649",
"rollout DB archive mismatches": "0",
"rollout DB archived files": "0",
"rollout DB archived rows": "0",
"rollout DB duplicate DB paths": "0",
"rollout DB duplicate rollout thread ids": "0",
"rollout DB malformed file names": "0",
"rollout DB missing active rows": "0",
"rollout DB missing archived rows": "0",
"rollout DB model providers": "openai=649",
"rollout DB rows": "649",
"rollout DB scan cap reached": "false",
"rollout DB scan errors": "0",
"rollout DB sources": "cli=489, subagent:thread_spawn=146, subagent:other=14",
"rollout DB stale rows": "0"
},
"remediation": null,
"durationMs": 394
},
"system.disk": {
"id": "system.disk",
"category": "disk",
"status": "ok",
"summary": "sufficient free disk space (106.6 GiB)",
"details": {
"CODEX_HOME available": "106.6 GiB",
"failure threshold": "1.0 GiB",
"warning threshold": "5.0 GiB",
"worktree available": "106.6 GiB"
},
"remediation": null,
"durationMs": 0
},
"system.environment": {
"id": "system.environment",
"category": "system",
"status": "ok",
"summary": "OS language C",
"details": {
"EDITOR": "not set",
"LANG": "C.UTF-8",
"VISUAL": "not set",
"os": "Ubuntu 26.4.0 (resolute) [64-bit]",
"os language": "C",
"os type": "Ubuntu",
"os version": "26.4.0"
},
"remediation": null,
"durationMs": 28
},
"terminal.env": {
"id": "terminal.env",
"category": "terminal",
"status": "ok",
"summary": "terminal metadata was detected",
"details": {
"COLORTERM": "truecolor",
"SSH_CLIENT": "present",
"SSH_CONNECTION": "present",
"SSH_TTY": "present",
"TERM": "xterm-256color",
"color output": "enabled",
"effective locale": "C.UTF-8",
"stderr is terminal": "true",
"stdin is terminal": "true",
"stdout is terminal": "true",
"terminal": "unknown",
"terminal size": "169x59"
},
"remediation": null,
"durationMs": 0
},
"terminal.title": {
"id": "terminal.title",
"category": "title",
"status": "ok",
"summary": "terminal title default",
"details": {
"terminal title activity": "true",
"terminal title items": "activity, project-name",
"terminal title project source": "cwd",
"terminal title project value": "gweeteve",
"terminal title source": "default"
},
"remediation": null,
"durationMs": 0
},
"updates.status": {
"id": "updates.status",
"category": "updates",
"status": "ok",
"summary": "update configuration is locally consistent",
"details": {
"cached latest version": "0.149.0",
"check for update on startup": "true",
"last checked at": "2026-08-21T06:52:30.036839330Z",
"latest version": "0.149.0",
"latest version status": "current version is not older",
"npm update target": "/home/gweeteve/.nvm/versions/node/v24.15.0/lib/node_modules/@openai/codex",
"update action": "npm install -g @openai/codex",
"version cache": "/home/gweeteve/.codex/version.json"
},
"remediation": null,
"durationMs": 158
}
}
}What issue are you seeing?
<html>
<body>
<!--StartFragment--><h1 data-line="0" class="code-line" dir="auto" id="support-report-unintended-fast-mode-billing-on-codex-subagents" style="margin-top: 0px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 2em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Support Report: Unintended Fast-Mode Billing on Codex Subagents</h1><p data-line="2" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><strong>Suggested ticket subject:</strong><span> </span>Codex subagents used Fast mode after I disabled it, consuming approximately 1,200 additional credits</p><h2 data-line="4" class="code-line" dir="auto" id="customer-and-environment" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Customer and environment</h2><ul data-line="6" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 0.7em; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><li data-line="6" class="code-line" dir="auto"><strong>Account email:</strong><span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">[ADD ACCOUNT EMAIL]</code></li><li data-line="7" class="code-line" dir="auto"><strong>Subscription:</strong><span> </span>ChatGPT Pro 5x</li><li data-line="8" class="code-line" dir="auto"><strong>Incident date:</strong><span> </span>August 20, 2026</li><li data-line="9" class="code-line" dir="auto"><strong>Timezone:</strong><span> </span>Europe/Paris (CEST, UTC+2)</li><li data-line="10" class="code-line" dir="auto"><strong>Codex client:</strong><span> </span>Codex TUI 0.148.0 on Linux</li><li data-line="11" class="code-line" dir="auto"><strong>Primary Codex thread ID:</strong><span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">01a01eec-5ce2-7050-8caf-2395f446144b</code></li><li data-line="12" class="code-line" dir="auto"><strong>Weekly allowance reset:</strong><span> </span>approximately 15:04 CEST (13:04 UTC)</li><li data-line="13" class="code-line" dir="auto"><strong>Observed impact:</strong><span> </span>almost 50% of the newly reset weekly allowance was consumed in less than five hours</li></ul><h2 data-line="15" class="code-line" dir="auto" id="executive-summary" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Executive summary</h2><p data-line="17" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">I am requesting a server-side billing audit and a credit adjustment for Codex usage on August 20, 2026.</p><p data-line="19" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">I disabled Fast mode on the primary Codex thread at<span> </span><strong>15:42:54 CEST</strong>. The primary thread remained on the Standard/default service tier afterward, and I did not issue another<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">/fast</code><span> </span>command. However, the Codex multi-agent runtime subsequently created at least<span> </span><strong>11 subagent threads</strong><span> </span>with the internal service tier explicitly set to<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">priority</code>, which is the Fast-mode tier recorded by this client version.</p><p data-line="21" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Those 11 subagents consumed<span> </span><strong>41
consommation_tokens_2026-08-20_par_modele_10min.csv
,113,684 tokens</strong>. Using the published GPT-5.6 Sol credit rates and the published 2.5x Fast multiplier:</p><ul data-line="23" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 0.7em; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><li data-line="23" class="code-line" dir="auto">Estimated cost in Standard mode:<span> </span><strong>799.764 credits</strong></li><li data-line="24" class="code-line" dir="auto">Estimated cost in Fast mode:<span> </span><strong>1,999.410 credits</strong></li><li data-line="25" class="code-line" dir="auto">Estimated unintended Fast-mode premium:<span> </span><strong>1,199.646 credits</strong></li></ul><p data-line="27" class="code-line code-active-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">The evidence does not indicate that Fast mode was globally re-enabled by me. It indicates that Fast/priority was applied independently when the child threads were created, despite the parent remaining on<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">default</code>.</p><h2 data-line="29" class="code-line" dir="auto" id="requested-resolution" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Requested resolution</h2><p data-line="31" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Please:</p><ol data-line="33" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 0.7em; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><li data-line="33" class="code-line" dir="auto">Audit the authoritative billing ledger for the primary thread and the 11 child thread IDs listed below.</li><li data-line="34" class="code-line" dir="auto">Confirm whether those child requests were billed with the GPT-5.6 Fast multiplier.</li><li data-line="35" class="code-line" dir="auto">Explain why child threads received<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">service_tier: "priority"</code><span> </span>after the parent thread had switched to<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">service_tier: "default"</code>.</li><li data-line="36" class="code-line" dir="auto">Restore the credits attributable to the unintended Fast multiplier. The local estimate is<span> </span><strong>approximately 1,200 credits</strong>, but the server-side ledger should determine the exact adjustment.</li><li data-line="37" class="code-line" dir="auto">Confirm whether other subagents without a locally visible service-tier marker were billed as Standard or Fast.</li><li data-line="38" class="code-line" dir="auto">Advise whether there is a setting that guarantees subagents cannot use Fast mode when the parent is in Standard mode.</li></ol><h2 data-line="40" class="code-line" dir="auto" id="incident-timeline" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Incident timeline</h2><p data-line="42" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">All local times below are CEST (UTC+2).</p>
Time | Event
-- | --
15:04 | Weekly allowance reset. This is the beginning of the measured five-hour window.
15:42:54 | The primary thread recorded service_tier: "default" for GPT-5.6 Sol.
16:44:08 | First confirmed post-disable child thread recorded service_tier: "priority".
16:44–17:23 | Eleven newly created subagents recorded the priority tier.
After 15:42 | No user /fast command was found in the primary-thread event log.
18:00, 18:12, 19:37 and 19:40 | The primary thread recorded service_tier: "default" again, confirming that the parent itself remained Standard.
20:04 | End of the five-hour analysis window.
<h2 data-line="151" class="code-line" dir="auto" id="evidence-that-this-was-not-a-user-reactivation" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Evidence that this was not a user reactivation</h2><ol data-line="153" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 0.7em; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><li data-line="153" class="code-line" dir="auto">The primary thread recorded<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">service_tier: "default"</code><span> </span>at<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">2026-08-20T13:42:54.773Z</code><span> </span>(15:42:54 CEST).</li><li data-line="154" class="code-line" dir="auto">No subsequent user message containing<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">/fast</code><span> </span>was found in the primary thread.</li><li data-line="155" class="code-line" dir="auto">The primary thread repeatedly recorded<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">service_tier: "default"</code><span> </span>later in the same session.</li><li data-line="156" class="code-line" dir="auto">The first affected child recorded<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">service_tier: "priority"</code><span> </span>at<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">2026-08-20T14:44:08.114Z</code>, over one hour after Fast was disabled on the parent.</li><li data-line="157" class="code-line" dir="auto">The child was created by a multi-agent<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">spawn_agent</code><span> </span>call. The spawn request did not contain a user-selected service-tier parameter.</li><li data-line="158" class="code-line" dir="auto">Ten additional child threads independently recorded<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">service_tier: "priority"</code><span> </span>through 17:23 CEST.</li><li data-line="159" class="code-line" dir="auto">The current<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">config.toml</code><span> </span>explicitly sets<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">service_tier = "default"</code>.</li></ol><p data-line="161" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">This points to a per-child tier assignment by the multi-agent runtime, rather than a global user-initiated Fast-mode reactivation. I am not alleging account compromise; I am asking OpenAI to investigate a likely product/runtime behavior or defect.</p><h2 data-line="163" class="code-line" dir="auto" id="limitations-and-confidence" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Limitations and confidence</h2><ul data-line="165" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 0.7em; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><li data-line="165" class="code-line" dir="auto">Local logs record token usage and service-tier events, but they do not contain OpenAI's authoritative charged-credit ledger.</li><li data-line="166" class="code-line" dir="auto">The exact plan allowance behind the percentage displayed in the UI is not present in the local logs.</li><li data-line="167" class="code-line" dir="auto">Service tier is not explicitly logged for every subagent. Unmarked activity was treated as Standard in the conservative estimate.</li><li data-line="168" class="code-line" dir="auto">The internal<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">priority</code><span> </span>label is the marker this Codex client recorded on the child threads associated with Fast behavior. OpenAI should confirm its server-side billing meaning for these requests.</li><li data-line="169" class="code-line" dir="auto">Token totals have high confidence because two independent aggregation methods matched exactly.</li><li data-line="170" class="code-line" dir="auto">The estimated 1,199.646-credit Fast premium has high confidence if OpenAI confirms that these 11<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">priority</code><span> </span>child threads were billed with the documented GPT-5.6 Fast multiplier.</li></ul><h2 data-line="172" class="code-line" dir="auto" id="files-available-for-secure-upload" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Files available for secure upload</h2><ul data-line="174" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 0.7em; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><li data-line="174" class="code-line" dir="auto"><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">consommation_tokens_2026-08-20_par_modele_10min.csv</code><span> </span>— full-day token usage grouped by model and ten-minute interval.</li><li data-line="175" class="code-line" dir="auto">Primary thread JSONL log for thread<span> </span><code style="font-family: Consolas, "Courier New", monospace; color: rgb(140, 140, 140); background-color: rgb(38, 38, 38); padding: 1px 3px; border-radius: 4px; font-size: 1em; line-height: 1.357em;">01a01eec-5ce2-7050-8caf-2395f446144b</code>.</li><li data-line="176" class="code-line" dir="auto">Eleven child-thread JSONL logs corresponding to the IDs above.</li></ul><p data-line="178" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 16px; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">The JSONL files may contain prompts, tool outputs, source-code context, and local paths. Please provide a secure upload channel if the raw logs are required. Redacted excerpts containing only timestamps, thread IDs, tier settings, and token counters can also be supplied.</p><h2 data-line="180" class="code-line" dir="auto" id="official-references" style="margin-top: 24px; font-weight: 600; margin-bottom: 16px; line-height: 1.25; font-size: 1.5em; padding-bottom: 0.3em; border-bottom: 1px solid rgba(255, 255, 255, 0.18); border-top-color: rgba(255, 255, 255, 0.18); border-right-color: rgba(255, 255, 255, 0.18); border-left-color: rgba(255, 255, 255, 0.18); color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;">Official references</h2><ul data-line="182" class="code-line" dir="auto" style="margin-top: 0px; margin-bottom: 0.7em; color: rgb(187, 190, 191); font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", system-ui, Ubuntu, "Droid Sans", sans-serif; font-size: 14px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; white-space: normal; text-decoration-thickness: initial; text-decoration-style: initial; text-decoration-color: initial;"><li data-line="182" class="code-line" dir="auto"><a href="https://learn.chatgpt.com/docs/pricing" data-href="https://learn.chatgpt.com/docs/pricing" style="color: rgb(72, 160, 199); text-decoration: none;">ChatGPT Learn — Codex pricing and credit rate card</a></li><li data-line="183" class="code-line" dir="auto"><a href="https://learn.chatgpt.com/docs/agent-configuration/speed" data-href="https://learn.chatgpt.com/docs/agent-configuration/speed" style="color: rgb(72, 160, 199); text-decoration: none;">ChatGPT Learn — Fast mode and credit multipliers</a></li><li data-line="184" class="code-line" dir="auto"><a href="https://developers.openai.com/api/docs/models/gpt-5.6-sol" data-href="https://developers.openai.com/api/docs/models/gpt-5.6-sol" style="color: rgb(72, 160, 199); text-decoration: none;">OpenAI Developers — GPT-5.6 Sol token pricing and long-context threshold</a></li></ul><!--EndFragment-->
</body>
</html>Support Report: Unintended Fast-Mode Billing on Codex Subagents
Suggested ticket subject: Codex subagents used Fast mode after I disabled it, consuming approximately 1,200 additional credits
Customer and environment
Account email: [ADD ACCOUNT EMAIL]
Subscription: ChatGPT Pro 5x
Incident date: August 20, 2026
Timezone: Europe/Paris (CEST, UTC+2)
Codex client: Codex TUI 0.148.0 on Linux
Primary Codex thread ID: 01a01eec-5ce2-7050-8caf-2395f446144b
Weekly allowance reset: approximately 15:04 CEST (13:04 UTC)
Observed impact: almost 50% of the newly reset weekly allowance was consumed in less than five hours
Executive summary
I am requesting a server-side billing audit and a credit adjustment for Codex usage on August 20, 2026.
I disabled Fast mode on the primary Codex thread at 15:42:54 CEST. The primary thread remained on the Standard/default service tier afterward, and I did not issue another /fast command. However, the Codex multi-agent runtime subsequently created at least 11 subagent threads with the internal service tier explicitly set to priority, which is the Fast-mode tier recorded by this client version.
Those 11 subagents consumed 41,113,684 tokens. Using the published GPT-5.6 Sol credit rates and the published 2.5x Fast multiplier:
Estimated cost in Standard mode: 799.764 credits
Estimated cost in Fast mode: 1,999.410 credits
Estimated unintended Fast-mode premium: 1,199.646 credits
The evidence does not indicate that Fast mode was globally re-enabled by me. It indicates that Fast/priority was applied independently when the child threads were created, despite the parent remaining on default.
Requested resolution
Please:
Audit the authoritative billing ledger for the primary thread and the 11 child thread IDs listed below.
Confirm whether those child requests were billed with the GPT-5.6 Fast multiplier.
Explain why child threads received service_tier: "priority" after the parent thread had switched to service_tier: "default".
Restore the credits attributable to the unintended Fast multiplier. The local estimate is approximately 1,200 credits, but the server-side ledger should determine the exact adjustment.
Confirm whether other subagents without a locally visible service-tier marker were billed as Standard or Fast.
Advise whether there is a setting that guarantees subagents cannot use Fast mode when the parent is in Standard mode.
Incident timeline
All local times below are CEST (UTC+2).
Time Event
15:04 Weekly allowance reset. This is the beginning of the measured five-hour window.
15:42:54 The primary thread recorded service_tier: "default" for GPT-5.6 Sol.
16:44:08 First confirmed post-disable child thread recorded service_tier: "priority".
16:44–17:23 Eleven newly created subagents recorded the priority tier.
After 15:42 No user /fast command was found in the primary-thread event log.
18:00, 18:12, 19:37 and 19:40 The primary thread recorded service_tier: "default" again, confirming that the parent itself remained Standard.
20:04 End of the five-hour analysis window.
The current local configuration also contains:
service_tier = "default"
Measured usage from 15:04 to 20:04 CEST
The token totals were reconstructed from Codex JSONL session logs by taking deltas of each thread's cumulative total_token_usage counters. Repeated snapshots were excluded, and counter resets were handled per thread. An independent sum of each request's last_token_usage produced the same totals exactly.
Usage by model
Model Input tokens Cached input tokens Output tokens Total tokens
GPT-5.6 Sol 239,343,621 233,368,832 767,134 240,110,755
GPT-5.6 Terra 5,802,408 5,486,848 35,729 5,838,137
Codex auto-review 96,720 41,728 913 97,633
Total 245,242,749 238,897,408 803,776 246,046,525
Additional observations:
2,215 model-usage events were recorded in the five-hour window.
48 distinct subagent rollouts consumed tokens.
Subagents accounted for approximately 75.7% of the minimum estimated credit consumption.
Cached input represented approximately 97.4% of all input tokens. Cached tokens are discounted but still consume credits.
Median request input was 102,891 tokens; the 95th percentile was 202,329; the maximum was 239,798.
No request exceeded 272,000 input tokens, so the published long-context threshold does not appear to explain the incident.
Credit calculation
The calculation uses the official Codex rate card published at ChatGPT Learn — Pricing:
GPT-5.6 Sol uncached input: 125 credits per 1M tokens
GPT-5.6 Sol cached input: 12.5 credits per 1M tokens
GPT-5.6 Sol output: 750 credits per 1M tokens
The official ChatGPT Learn — Speed documentation states that GPT-5.6 Fast mode consumes credits at 2.5x the Standard rate.
Because the local input_tokens counter includes cached input, uncached input was calculated as:
uncached input = input tokens - cached input tokens
For GPT-5.6 Sol:
Standard credits =
uncached input × 125 / 1,000,000
- cached input × 12.5 / 1,000,000
- output × 750 / 1,000,000
Fast credits = Standard credits × 2.5
Estimated credit breakdown
Scope and observed tier Total tokens Estimated credits Notes
Primary thread, Fast before it was disabled 7,383,937 269.525 Would have been 107.810 in Standard; premium 161.715
Primary thread, confirmed Standard 74,742,947 1,108.298 Parent remained Standard after 15:42
11 subagents, confirmed Fast after disable 41,113,684 1,999.410 Would have been 799.764 in Standard; premium 1,199.646
Subagents without a visible tier marker, GPT-5.6 Sol 116,870,187 2,223.437 Conservatively estimated as Standard; server confirmation required
Terra and auto-review activity 5,985,770 57.972 Minor contribution
Minimum supported total estimate 246,046,525 5,658.642 Treats unmarked requests as Standard
If every locally unmarked GPT-5.6 subagent request had also been billed as Fast, the upper estimate would be approximately 9,074.694 credits. There is not enough local evidence to assert that this happened, which is why the report uses 5,658.642 credits as the conservative estimate and requests a server-side tier audit.
Counterfactual impact
If the 11 post-disable Fast subagents had run in Standard mode, the estimated total would have been approximately:
5,658.642 - 1,199.646 = 4,458.996 credits
Therefore, the unintended Fast multiplier increased the conservative five-hour estimate by approximately 21.2%.
If the locally estimated 5,658.642 credits correspond to the roughly 50% weekly-allowance decrease shown in the UI, the unintended Fast premium alone represents approximately 10.6 percentage points of the weekly allowance. This percentage is an inference; only OpenAI's authoritative usage ledger can confirm the exact relationship.
Confirmed post-disable Fast subagents
Fast applied (CEST) Child thread ID Agent path Tokens Standard credits Fast credits Premium
16:44:08 01a01fa1-1e96-7152-819e-c494551f0e3f /root/rereview_archive_hardening 938,769 33.163 82.908 49.745
16:44:44 01a01fa1-ab3a-7521-a2b3-0f5ab87c3543 /root/implement_direct_reply_feedback 15,073,587 248.996 622.491 373.494
16:53:31 01a01fa9-b635-7c92-a273-be97e762db19 /root/audit_autonomy_signal_fix 2,445,438 51.808 129.519 77.711
17:01:39 01a01fb1-28e4-75e3-ae2e-6928b82d7d29 /root/review_direct_reply_feedback 3,127,325 63.402 158.506 95.104
17:02:18 01a01fb1-c25b-72e2-afa0-733c6c414fa3 /root/implement_autonomy_signals 6,190,102 121.377 303.441 182.065
17:04:33 01a01fb3-d264-7053-b649-cfb09512d740 /root/audit_full_dialogue_engagements 5,144,970 98.857 247.142 148.285
17:08:20 01a01fb7-46b8-7512-937b-6b308c8e820e /root/review_autonomy_signals 2,664,040 53.011 132.528 79.517
17:13:01 01a01fbb-902c-7d31-96c3-b00eb307dd94 /root/review_full_dialogue_engagements 432,672 16.163 40.408 24.245
17:13:44 01a01fbc-3ae1-75b1-9e1e-51c152d49092 /root/rereview_direct_reply_fix 2,441,727 50.243 125.606 75.364
17:17:49 01a01fbf-f8af-7a51-9db1-33365d5cb64e /root/rereview_autonomy_unknowns 1,699,460 38.534 96.335 57.801
17:23:27 01a01fc5-1df5-79b0-babd-cf49740e0917 /root/rereview_strict_trust_sources 955,594 24.210 60.526 36.315
Total 41,113,684 799.764 1,999.410 1,199.646
Evidence that this was not a user reactivation
The primary thread recorded service_tier: "default" at 2026-08-20T13:42:54.773Z (15:42:54 CEST).
No subsequent user message containing /fast was found in the primary thread.
The primary thread repeatedly recorded service_tier: "default" later in the same session.
The first affected child recorded service_tier: "priority" at 2026-08-20T14:44:08.114Z, over one hour after Fast was disabled on the parent.
The child was created by a multi-agent spawn_agent call. The spawn request did not contain a user-selected service-tier parameter.
Ten additional child threads independently recorded service_tier: "priority" through 17:23 CEST.
The current config.toml explicitly sets service_tier = "default".
This points to a per-child tier assignment by the multi-agent runtime, rather than a global user-initiated Fast-mode reactivation. I am not alleging account compromise; I am asking OpenAI to investigate a likely product/runtime behavior or defect.
Limitations and confidence
Local logs record token usage and service-tier events, but they do not contain OpenAI's authoritative charged-credit ledger.
The exact plan allowance behind the percentage displayed in the UI is not present in the local logs.
Service tier is not explicitly logged for every subagent. Unmarked activity was treated as Standard in the conservative estimate.
The internal priority label is the marker this Codex client recorded on the child threads associated with Fast behavior. OpenAI should confirm its server-side billing meaning for these requests.
Token totals have high confidence because two independent aggregation methods matched exactly.
The estimated 1,199.646-credit Fast premium has high confidence if OpenAI confirms that these 11 priority child threads were billed with the documented GPT-5.6 Fast multiplier.
Files available for secure upload
consommation_tokens_2026-08-20_par_modele_10min.csv — full-day token usage grouped by model and ten-minute interval.
Primary thread JSONL log for thread 01a01eec-5ce2-7050-8caf-2395f446144b.
Eleven child-thread JSONL logs corresponding to the IDs above.
The JSONL files may contain prompts, tool outputs, source-code context, and local paths. Please provide a secure upload channel if the raw logs are required. Redacted excerpts containing only timestamps, thread IDs, tier settings, and token counters can also be supplied.
Official references
ChatGPT Learn — Codex pricing and credit rate card
ChatGPT Learn — Fast mode and credit multipliers
OpenAI Developers — GPT-5.6 Sol token pricing and long-context threshold
What steps can reproduce the bug?
Uploaded thread: 01a02317-b3a5-73f0-8fdc-78b3e992549f
What is the expected behavior?
_No response_
Additional information
_No response_
1 Comment
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action