OpenAI models

Quick choice

Isolated and utility completions on the Codex runtime use the same account routes as the model picker. A model listed by your ChatGPT account does not require an OpenAI API key for these completions. An explicit auth profile or utility-model override still controls the request; OpenClaw does not replace it with another account.

Retired subscription model references

GPT-5.4 and GPT-5.4 Mini are retired from the ChatGPT-account Codex route. Run openclaw doctor --fix to replace persisted subscription references with their documented successors: openai/gpt-5.6-terra and openai/gpt-5.6-luna, respectively. This includes defaults, per-agent model selections, automation overrides, and unlocked session overrides whose selected route is known. The Platform API-key route is unaffected. Doctor retains pinned overrides when their successor is outside the agent's model policy, or when clearing an override would keep the same retired model and account. Doctor also retains the original reference and warns when its declared successor is retired or definitively unsupported on the selected account route; unknown availability and temporary cooldowns do not block migration. It reports the model or policy change needed, along with unresolved or conflicting account routes. Review the repair output, restart the Gateway, and re-enable any automation that was disabled after repeated failures.

Models your API key lists

With an OpenAI API key, openclaw models list --provider openai shows every chat model your account's /v1/models lists that runs on the Responses API, including models newer than your OpenClaw release. Models without a catalog row use conservative defaults: text input, a 128k context window, and unknown cost. Embedding, image, video, audio, realtime, speech, transcription, moderation, and legacy completions models stay hidden, as do models past their OpenAI shutdown date. Search models stay hidden because they run only on Chat Completions; live, unrecognized cyber, experiment, and alpha IDs stay hidden because they can fail on Responses. The documented gpt-5.6-cyber Daybreak model is included when the account lists it; approved Daybreak Red access is still required. GPT-3.5, GPT-4 (including GPT-4o and GPT-4.1), o1, o3, and o4 models stay hidden from the list because their tool capabilities differ from the modern models. You can still select one by its reference, for example as agents.defaults.model.primary or a session model. OpenClaw automatically uses its function-tool runtime instead of Codex's native custom tools; no manual agentRuntime.id override is needed. GPT-5-family models keep the Codex runtime. o1, o3-mini, GPT-3.5, GPT-4, GPT-4 Turbo, and GPT-4.1 nano also reject OpenAI's hosted web search, which the OpenClaw runtime adds by default: set tools.web.search.provider to a managed provider such as brave, or tools.web.search.enabled: false, before using them (see Native OpenAI web search).

Daybreak alias retirement

OpenAI retires openai/gpt-daybreak-blue-latest and openai/gpt-daybreak-red-latest on October 28, 2026. Only the aliases retire; the underlying models remain available to approved accounts. OpenClaw marks the aliases deprecated and stops choosing Blue as an implicit cyber failover target. Explicit alias references remain unchanged; this is not an automatic switch to a different model.

Before the deadline:

1. Check your Daybreak settings and the specific model IDs available under your approved access. Do not infer a replacement from Blue/Red or another account's catalog. Listing a model is not proof of Daybreak entitlement.
2. Run openclaw doctor --lint --only openai/daybreak-alias-retirement to locate configured alias references. Replace them with your supported specific model ID, preserving the intended auth profile and runtime. This advisory check has no automatic repair: it does not block updates, guess a replacement, or clear your selection. Review primary/fallback models, model-map keys, per-agent and tool selections, saved session overrides, and scheduled jobs as applicable. Use /model to change an existing conversation's model.
3. For Responses API use, follow the Daybreak guide, including its account settings and request requirements. Preserve explicitly configured reasoning capabilities and service tiers for the specific model; the old aliases' capabilities are not a contract for every underlying model.
4. Select an explicit access program in the chat Access control, or use Codex's own Daybreak controls with an approved specific model. Selecting a model or access program does not grant account access.

Automatic cyber recovery retries a replay-safe cyber refusal once on the same model and account with authorized Blue Access. It preserves the runtime, effort, speed, and saved session settings; ordinary configured model fallback remains separate. Run openclaw doctor --fix to remove the retired cyberFailover.model field from agents.defaults.embeddedAgent.cyberFailover (OpenClaw runtime) or plugins.entries.codex.config.appServer.cyberFailover (native Codex). Doctor preserves mode and cooloffMs, including explicit mode: "off". See model failover and Codex runtime behavior.

Chat Access control

Open the chat effort/speed menu. Access appears above Effort and Speed only when the selected account advertises at least one Daybreak program for that model and runtime. Unavailable programs are hidden, not disabled. Missing or unverified entitlement metadata hides the control.

• Standard and the available Blue or Red options explicitly select the requested access program.
• Use model default clears the session override. This is different from selecting Standard: the configured model or provider may choose a default program.
• If account capabilities no longer expose Access, open the model menu and choose Clear saved Access to restore the model default. The saved choice is never cleared automatically, and unavailable programs remain hidden.

The session override is saved independently of speed, survives a session reset, and is inherited by a fork unless replaced. Changing the selection affects the next admitted turn, not an already-running model call.

The OpenClaw runtime sends the program to its official Responses endpoint. Native Codex uses its supported turn-scoped access-program parameter. This OpenClaw integration currently supports explicit native programs only with ChatGPT subscription authentication and rejects API-key selections before submission. Upstream Codex can forward programs when its separate API-key feature is enabled; this integration does not support that mode. Reset Access to the model default or use the OpenClaw runtime with an explicit model parameter for Platform API-key requests. Native Codex checks its authenticated model/list access-program metadata before submitting a Daybreak turn. A stale saved selection fails visibly before submission.

Entitlement comes only from authenticated available_access_programs.cyber metadata (or native Codex's availableAccessPrograms.cyber), scoped to the selected account, model, route, and runtime. Model listing membership or subscription type alone is not entitlement. For Platform API keys, OpenClaw reads explicit program metadata from the same authenticated Codex model endpoint used by the official Codex client. The public /v1/models response remains the model-availability authority: a capability row cannot add an unlisted model. The same API key scopes both reads to its organization and project; custom endpoints and request-header overrides do not inherit these grants.

This discovery uses read-only requests, not billable Responses probes. Successful Platform and ChatGPT HTTP observations are cached for up to 60 seconds by credential and endpoint; explicit catalog refresh bypasses that cache. This HTTP TTL does not periodically refresh an already published account catalog; its normal refresh and credential-change lifecycle still applies. A replaced credential, missing or empty program metadata, or a failed capability refresh never reuses another account's grants. Unknown or unavailable programs stay hidden, without removing ordinary model access. Provider-side access changes may also have their own propagation delay. Native Codex owns its separate model-list cache and account-change invalidation. Explicit model parameters remain available; the provider enforces entitlement on every actual request.

Responses access program and speed

For the OpenClaw runtime on an official OpenAI Platform or ChatGPT subscription Responses endpoint, set the selected model's params.accessPrograms.cyber to standard, daybreak_blue, or daybreak_red. Omit it or use accessPrograms: {} to retain OpenAI's model/account/project defaults. Neither accessPrograms nor cyber accepts null. Native Codex also accepts this model setting on ChatGPT authentication, with the session override taking precedence. This setting never grants access. An explicit incompatible program or missing entitlement returns the provider error; OpenClaw does not silently switch to standard safeguards.

Daybreak and speed are separate request settings. For example, an approved account using the publicly documented GPT-6.1 Sol route can request:

{  agents: {    defaults: {      models: {        "openai/gpt-6.1-sol": {          agentRuntime: { id: "openclaw" },          params: {            accessPrograms: { cyber: "daybreak_blue" },            serviceTier: "ultrafast", // "priority" for Fast          },        },      },    },  },}

GPT-6.1 Sol and GPT-6 Astra require Daybreak Red approval and project access even though their request value is daybreak_blue; daybreak_red is not the correct value for those models. Match the program to the selected model in the Daybreak guide. The model and account must independently support the requested Fast or Ultrafast tier. OpenClaw sends both settings without applying the retiring aliases' speed limits to specific model IDs. Provider authorization and fulfillment remain server-owned.

Until retirement, explicitly selected aliases retain their existing Responses metadata: Blue keeps reasoning enabled and supports Standard/Fast; Red also supports reasoning off and uses Standard. Neither supports Ultrafast. Their unknown cost estimates do not mean requests are free. These legacy rows stay only to keep existing selections interpretable during migration, not to name a replacement or grant access.

GPT-6 Astra

Select openai/gpt-6-astra with an OpenAI API-key profile or a ChatGPT/Codex subscription that has access to Astra. Access is rolling out; a successful account catalog remains authoritative, so adding model support does not grant access to an account that has not received it. If ChatGPT/Codex catalog discovery is unavailable, the offline fallback list omits Astra until account discovery succeeds.

openclaw models set openai/gpt-6-astra

Astra uses the Responses API for agent tool calls. It supports text and image input, a 1,050,000-token context window, and up to 128,000 output tokens. OpenClaw retains its ordinary 272,000-token active input budget by default. The supported reasoning efforts are low, medium, high, xhigh, and max. OpenClaw defaults Astra to medium on both the OpenClaw and Codex runtimes when the account supports that effort. The OpenAI provider owns this default, so model selection, Control UI, and Codex turn requests share it. Explicit agent, model, global, and session thinking settings still take precedence; switching models does not clear an existing high override. Use /think default to clear a session override. An existing minimal setting maps to low. Astra cannot disable reasoning; off never sends the unsupported none effort. Temperature and top_p are not sent. These defaults also apply to configured Astra model entries without explicit reasoning or temperature compatibility metadata. Azure Responses deployments continue to use their configured capabilities.

/think ultra is also available on the OpenClaw and Codex runtimes. Ultra enables proactive sub-agent orchestration; it is not a raw Responses API effort. OpenClaw uses max, while native Codex selects Astra's model-defined effort (xhigh).

Standard pricing per million tokens is $10 input, $1 cache reads, $12.50 cache writes, and $50 output. Requests above 272K input tokens have higher rates. See the Astra model reference and migration guide.

Async tools, steering, and reasoning changes

Use an OpenAI Platform API-key profile and the built-in OpenClaw runtime for these Astra capabilities. They require the official https://api.openai.com/v1 Responses endpoint. Configure the existing model settings:

{  agents: {    defaults: {      models: {        "openai/gpt-6-astra": {          agentRuntime: { id: "openclaw" },          params: {            transport: "auto",            responsesServerCompaction: false,          },        },      },    },  },}

• Async function calls: Astra can continue reasoning while OpenClaw runs a direct function tool. OpenClaw sends the completed result in the next model request after the active response finishes. This applies to direct tools; code-mode tools retain their existing execution flow. sessions_yield stays synchronous: it is how the model waits, so the response pauses there and the next request delivers the earlier async results.
• Mid-turn steering: Steering messages can reach Astra while it is reasoning, using the active session's cached WebSocket. Use auto or websocket-cached; SSE keeps ordinary queued steering at the next available runtime boundary. Each live batch owns one response; later messages can steer its successor. Context or payload hooks that rewrite the active request's prefix keep ordinary queued delivery.
• Reasoning changes without rebuilding the cached prefix: Change the thinking level, for example with /think high, before the next user turn. OpenClaw preserves the original request-level effort and places a configuration_update at the new turn. This optimization works across matching session history over SSE or cached WebSockets, including concurrent SSE requests in the same session. Automatic steering continuations keep their inherited settings. If steering waits for a tool result or approval, the explicit continuation uses current request settings, including output limits and reasoning settings, without repeating accepted steering. Earlier configuration_update items retain their effect; a changed request-level effort does not replace those controls. When accepted steering waits for a tool result or approval and its history contains effort controls, finish that input with a compatible Astra model and mode before switching.

The example disables automatic server compaction because OpenAI cannot combine it with configuration updates. Cache-preserving effort changes also exclude automatic truncation, pro mode, and API multi-agent mode. The original effort and admitted controls survive transport expiry and Gateway restarts in saved provider replay metadata. Matching history replays the same prefix without extending socket or provider cache lifetimes. Rewritten or compacted history, incompatible settings, or a changed model, route, session, or auth profile starts a fresh request using the selected effort. Older transcripts without this metadata also start fresh after transport expiry.

The native Codex harness owns its own Responses loop. OpenClaw enables its cache-preserving reasoning-effort updates for managed threads, including when native code mode is disabled. Codex uses its model catalog to decide eligibility and keeps unsupported models on ordinary request-level effort. This does not enable the built-in-runtime async-tool or steering implementation inside Codex.

GPT-6.1 Sol

Select openai/gpt-6.1-sol with an OpenAI API-key profile or a ChatGPT/Codex subscription that exposes it:

openclaw models set openai/gpt-6.1-sol

It uses the Responses API for tool calls, accepts text and images, and supports a 1,050,000-token context window with up to 128,000 output tokens. Direct OpenAI API-key traffic keeps its 272,000-token active input budget by default. The OpenClaw harness defaults to a 500,000-token active window with ChatGPT/Codex subscription auth; native Codex owns its context window and compaction. Explicit context budgets, other models, and the Astra setup default are unchanged. A successful account catalog remains authoritative, including smaller account-specific windows; subscription offline hints do not imply access.

Reasoning defaults to medium and supports low, medium, high, xhigh, and max. Unlike GPT-6 Sol, GPT-6.1 Sol cannot disable reasoning: off does not send none, and minimal maps to low. OpenClaw offers /think ultra through its existing orchestration mode; native Codex uses the account's advertised efforts.

Standard API prices per million tokens are $2 input, $0.10 cache reads, $2.50 cache writes, and $10 output. Above 272K input tokens, input and cache rates double and output costs 1.5 times the standard rate for the full request. See the GPT-6.1 Sol model reference for current capabilities, pricing, and regional availability.

GPT-6 Sol and Luna

Select openai/gpt-6-sol or openai/gpt-6-luna with an OpenAI API-key profile or a ChatGPT/Codex subscription that exposes the model. Check the selected account's catalog, then choose the model:

openclaw models list --provider openaiopenclaw models set openai/gpt-6-sol

Use openclaw models set openai/gpt-6-luna to select Luna. API organization and Codex workspace access can differ. A successful account catalog is authoritative; OpenClaw does not add subscription access or silently substitute another model. When subscription discovery is unavailable, the offline fallback list omits both models. Existing model selections stay unchanged, and fresh OpenAI setup continues to use Astra.

Both models use the Responses API for agent tool calls and support text and image input, a 1,050,000-token context window, and up to 128,000 output tokens. OpenClaw defaults to a 272,000-token active input budget. Native Codex follows the selected account's advertised context limits.

The API reasoning efforts are none, low, medium, high, xhigh, and max. OpenClaw defaults to medium when available; existing explicit thinking settings still take precedence. /think off selects none, and /think default clears a session override. Native Codex uses the effort levels reported by the selected account. OpenClaw's /think ultra mode uses max; native Codex offers Ultra only when the account advertises it.

Standard API pricing per million tokens:

See the GPT-6 Sol model reference and GPT-6 Luna model reference for current capabilities and pricing.

GPT-5.6 limited preview

OpenClaw recognizes the exact openai/gpt-5.6-sol, openai/gpt-5.6-terra, and openai/gpt-5.6-luna model ids. All three expose xhigh and max reasoning in the current catalog. OpenAI describes Sol as the flagship tier, Terra as the balanced tier, and Luna as the fast, lower-cost tier. See the GPT-5.6 launch announcement and access guide.

OpenAI's GPT-5.6 Sol model page documents the bare openai/gpt-5.6 id as a supported alias for Sol. Fresh API-key and ChatGPT/Codex OAuth setup use openai/gpt-6-astra. Existing GPT-5.6 selections retain their full Sol identity. Run openclaw doctor --fix to rewrite persisted bare OpenAI refs to that full identity. The native Codex catalog can show the exact Sol, Terra, and Luna ids depending on workspace access. Check the current account with:

openclaw models list --provider openai

API organization and Codex workspace access can differ. If GPT-5.6 is not available, select GPT-5.5 explicitly:

openclaw models set openai/gpt-5.5

OpenClaw surfaces the upstream access error and does not silently replace a GPT-5.6 selection with GPT-5.5.