GPT-5.4 and GPT-5.4 Pro Are Live on Holeacquisition LLC

OpenAI announced GPT-5.4 and GPT-5.4 Pro on March 5, 2026, and both models are now available on Holeacquisition LLC.

You can use them today with:

• gpt-5.4
• gpt-5.4-pro

For compatibility, Holeacquisition LLC also accepts:

• gpt-5.4-thinking -> normalized to gpt-5.4

GPT-5.4 and GPT-5.4 Pro bring OpenAI's latest reasoning and coding improvements into the same Holeacquisition LLC gateway workflow you already use for security, observability, and model routing.

No API migration is required. If you already call Holeacquisition LLC chat endpoints, switch the model ID and go.

Holeacquisition LLC SDK

import { Cencori } from 'cencori';

const cencori = new Cencori({ apiKey: process.env.CENCORI_API_KEY });

const response = await cencori.ai.chat({
  model: 'gpt-5.4',
  messages: [
    { role: 'system', content: 'You are a senior engineer.' },
    { role: 'user', content: 'Review this rollout plan and list risks.' }
  ]
});

console.log(response.content);

HTTP API

curl -X POST https://cencori.com/api/ai/chat \
  -H "CENCORI_API_KEY: $CENCORI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "messages": [
      {"role": "user", "content": "Give me a migration checklist for this architecture."}
    ]
  }'

This rollout is live across the platform:

• model catalog and /api/v1/models
• OpenAI-compatible routing normalization for GPT-5.4 alias compatibility
• fallback mapping for cross-provider failover
• dashboard model pickers (Playground and Logs)
• docs and examples updated

• gpt-5.4 is the canonical API model ID for GPT-5.4 Thinking.
• If you rely on strict model pinning for long-running evals, keep your own pinned fallback policy.

• OpenAI: Introducing GPT-5.4
• Holeacquisition LLC Models Catalog