OpenAI

Nexgile Code offers two distinct ways to run OpenAI models, and they appear as two separate entries in the provider dropdown:

• OpenAI — the native OpenAI API, billed per token against an API key from your OpenAI platform account.
• OpenAI - ChatGPT Plus/Pro — sign in with an existing ChatGPT subscription; usage is covered by the subscription instead of metered API billing.

This page covers both.

The “OpenAI” entry (API key)

Get an API key

1. Sign in at platform.openai.com.
2. Create a key on the API keys page — when no key is set, the settings panel shows a Get OpenAI API Key button that links directly to platform.openai.com/api-keys.
3. Make sure the platform account has billing enabled.

Configure in Nx IDE

1. Open Settings → Providers and pick OpenAI.
2. Paste the key into OpenAI API Key (stored in the editor’s Secret Storage).
3. Pick a model.

Optional controls:

• Use custom base URL — point at a compatible relay instead of https://api.openai.com/v1. For arbitrary third-party OpenAI-shaped servers, prefer the OpenAI Compatible provider instead.
• Service tier — for models that support it, choose Standard, Flex (cheaper, higher latency), or Priority (faster, more expensive). The dropdown only appears when the selected model offers those tiers.

Model families

The default model is gpt-5.1-codex-max. The catalog spans several families (exact roster in the model dropdown):

Family	Examples	Notes
GPT-5.5	gpt-5.5, gpt-5.5-pro	Newest general line; Pro is a compute-intensive reasoning tier
GPT-5.4	gpt-5.4, gpt-5.4-mini, gpt-5.4-nano, gpt-5.4-pro	Strong general models with Flex/Priority tiers and verbosity control
Codex (coding)	gpt-5.1-codex-max, gpt-5.3-codex, gpt-5.2-codex, gpt-5.1-codex, gpt-5.1-codex-mini, gpt-5-codex, codex-mini-latest	Optimized for long-horizon agentic coding
GPT-5.x general	gpt-5.2, gpt-5.1, gpt-5, gpt-5-mini, gpt-5-nano, gpt-5.2-chat-latest, gpt-5.3-chat-latest	Broad capability range, up to ~1M-token context on the larger models
GPT-4.1	gpt-4.1, gpt-4.1-mini, gpt-4.1-nano	Previous generation, large context
o-series (reasoning)	o3, o4-mini, o3-mini (each with high/low variants), o1-preview, o1-mini	Dedicated reasoning models
GPT-4o	gpt-4o, gpt-4o-mini	Legacy multimodal line

Behavior notes

• Reasoning effort: most GPT-5.x models accept effort levels from None up to Extra High; Codex models default higher (the default model ships at Extra High). See Model Behavior & Costs.
• Verbosity: the GPT-5.4/5.5 general models support the Output Verbosity control (Low / Medium / High).
• Temperature: the modern reasoning-centric models do not accept a temperature override; the control is hidden or inert for them by design.
• Prompt caching: supported across most of the catalog; cache reads bill at roughly a tenth of the input price. gpt-5.1-codex-max supports extended 24-hour cache retention.
• Image input: supported by the GPT-5.x, GPT-4.1, and GPT-4o families.

The “OpenAI - ChatGPT Plus/Pro” entry (subscription sign-in)

If you already pay for ChatGPT Plus or Pro, this entry runs Codex-line models against your subscription — no API key and no per-token charges inside Nx IDE.

How sign-in works

1. Open Settings → Providers and pick OpenAI - ChatGPT Plus/Pro.
2. Click Sign in to OpenAI Codex.
3. Your browser opens OpenAI’s authorization page (auth.openai.com). Approve access with the account that holds your ChatGPT subscription.
4. The browser redirects to a local callback (http://localhost:1455/auth/callback) that hands the token back to the editor. Port 1455 must be free on your machine during sign-in.

Tokens are refreshed automatically; if a request hits an authorization error, Nexgile Code force-refreshes the token once and retries before asking you to sign in again. A Sign out button appears in the same panel once connected.

What you get

• A subset of the Codex-era models: gpt-5.3-codex (default), gpt-5.3-codex-spark, gpt-5.1-codex-max, gpt-5.1-codex, gpt-5.2-codex, gpt-5.1-codex-mini, gpt-5-codex, gpt-5-codex-mini, plus gpt-5.5, gpt-5.4, gpt-5.4-mini, gpt-5.2, gpt-5.1, and gpt-5.
• Requests route to OpenAI’s Codex backend rather than the metered API, so the task-header cost readout stays at $0 — usage counts against your subscription’s limits instead.
• A rate-limit dashboard in the settings panel (visible once signed in) showing where you stand against those subscription limits.

Note

Subscription limits are enforced by OpenAI, not by Nx IDE. Long agentic sessions can exhaust an hourly or weekly allowance; the dashboard tells you before you hit the wall.

Choosing between the two entries

	OpenAI (API key)	OpenAI - ChatGPT Plus/Pro
Billing	Per token, pay-as-you-go	Covered by subscription
Model choice	Full catalog	Codex-focused subset
Limits	Account rate limits, your own cost caps	Subscription allowances
Best for	Teams, heavy use, model variety	Individuals already subscribed

Related

• Providers Overview
• Model Behavior & Costs
• OpenAI-Compatible Endpoints
• Connecting an AI Provider