diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 66f6f74..f84afe4 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -2,9 +2,8 @@ name: CI
 
 on:
   push:
-    branches: ['**']
+    branches: [main]
   pull_request:
-    branches: ['**']
 
 jobs:
   test-and-lint:
diff --git a/README.md b/README.md
index 77103fd..1094aa2 100644
--- a/README.md
+++ b/README.md
@@ -10,7 +10,7 @@ Main features:
 
 - `📝` Works directly with gettext `.po` files
 - `🤖` Translates only untranslated entries using AI
-- `🧠` Uses OpenAI `gpt-4o` by default for translation
+- `🧠` Uses OpenAI `gpt-5.4` by default for translation
 - `🏷️` Respects gettext context (`msgctxt`) when translating entries
 - `🔁` Supports singular and plural translations
 - `⚠️` Skips fuzzy entries by default
@@ -21,34 +21,12 @@ Main features:
 
 1. Read the `.po` file and parse its entries.
 2. Find entries with empty or missing translations.
-3. Send those strings to OpenAI `gpt-4o` for translation while preserving gettext context such as `msgctxt`.
+3. Send those strings to OpenAI `gpt-5.4` for translation while preserving gettext context such as `msgctxt`.
 4. Write the translated values back into the same `.po` file.
 
 The translation API uses OpenAI `json_schema` structured outputs. Only models that support `json_schema` structured outputs are valid for `msgai`.
 
-<details>
-<summary>Supported model families</summary>
-
-- `gpt-4o`
-- `gpt-4o-mini`
-- `gpt-4.1`
-- `gpt-4.1-mini`
-- `gpt-4.1-nano`
-- `gpt-5`
-- `gpt-5-mini`
-- `gpt-5-nano`
-- `gpt-5-pro`
-- `gpt-5.1`
-- `gpt-5.2`
-- `gpt-5-codex`
-- `gpt-5.1-codex`
-- `gpt-5.1-codex-mini`
-- `gpt-5.1-codex-max`
-- `gpt-5.2-codex`
-
-Dated snapshots are accepted where the model family supports them.
-
-</details>
+Any OpenAI model that supports `json_schema` structured outputs can be used via the `--model` flag.
 
 By default, entries marked as `fuzzy` are skipped. If you use `--include-fuzzy`, `msgai` will translate those entries too and remove the fuzzy flag after applying the result.
 
@@ -87,7 +65,7 @@ Options:
 - `--dry-run`: list untranslated `msgid` values only, with no API calls and no file changes
 - `--include-fuzzy`: include fuzzy entries for translation and clear their fuzzy flag after translation
 - `--source-lang LANG`: set the source language of `msgid` strings as an ISO 639-1 code such as `en` or `uk`
-- `--model MODEL`: set the OpenAI model used for translation; default is `gpt-4o`. Only models with `json_schema` structured outputs are supported.
+- `--model MODEL`: set the OpenAI model used for translation; default is `gpt-5.4`. Only models with `json_schema` structured outputs are supported.
 - `--api-key KEY`: pass the OpenAI API key directly instead of using `OPENAI_API_KEY`
 - `--fold-length N`: set PO line fold length when writing files. Use `0` to disable folding and minimize formatting-only diffs. Default: `0`
 - `--context TEXT`: additional instructions for the translation model in English, appended to the system prompt (e.g. "use formal tone", "don't translate currency names")
@@ -111,7 +89,7 @@ Example `msgai.config.yml`:
 
 ```yaml
 source-lang: en
-model: gpt-4o
+model: gpt-5.4
 include-fuzzy: false
 fold-length: 80
 context: "use formal tone"
diff --git a/src/cli/index.ts b/src/cli/index.ts
index 21b2765..689a3cf 100644
--- a/src/cli/index.ts
+++ b/src/cli/index.ts
@@ -47,7 +47,7 @@ function parseArgs(argv: string[]): CliArgs {
       })
       .option('model', {
         type: 'string',
-        description: 'OpenAI model to use for translation. Default: gpt-4o',
+        description: 'OpenAI model to use for translation. Default: gpt-5.4',
       })
       .option('fold-length', {
         type: 'number',
diff --git a/src/cli/runTranslate.ts b/src/cli/runTranslate.ts
index b701452..6c30c19 100644
--- a/src/cli/runTranslate.ts
+++ b/src/cli/runTranslate.ts
@@ -10,12 +10,7 @@ import {
   getLanguage,
   getPluralForms,
 } from '../po';
-import {
-  resolveApiKey,
-  translateStrings,
-  validateModel,
-  SUPPORTED_STRUCTURED_OUTPUT_MODELS,
-} from '../translate';
+import { resolveApiKey, translateStrings } from '../translate';
 import { validateSourceLang } from '../validate-source-lang';
 
 const TRANSLATE_BATCH_SIZE = 15;
@@ -47,6 +42,11 @@ function getApiErrorMessage(err: unknown): string | null {
         return `Quota exceeded (out of credits or usage limit). Check plan and billing: https://platform.openai.com/settings/organization/billing`;
       }
       return `Rate limit reached. Request was retried; if this persists, slow down or check https://developers.openai.com/api/docs/guides/rate-limits`;
+    case 400:
+      if (/response_format|json_schema|structured/i.test(message)) {
+        return `The specified model may not support json_schema structured outputs required by msgai. Try a compatible model like gpt-5.4. API error: ${message}`;
+      }
+      return `Invalid request: ${message}`;
     case 500:
       return `OpenAI server error. Retry later; see https://status.openai.com/`;
     case 503:
@@ -56,13 +56,6 @@ function getApiErrorMessage(err: unknown): string | null {
   }
 }
 
-function getInvalidModelMessage(model: string): string {
-  return [
-    `Invalid --model "${model}". msgai only supports OpenAI models with json_schema structured outputs.`,
-    `Supported model families: ${SUPPORTED_STRUCTURED_OUTPUT_MODELS.join(', ')}.`,
-  ].join(' ');
-}
-
 export type TranslateCommandArgs = {
   poFilePath?: string;
   dryRun: boolean;
@@ -91,7 +84,7 @@ export async function runTranslate(
     debugLogger.log('cli.runTranslate', 'Starting translation run', {
       poFilePath,
       sourceLang,
-      model: model ?? 'gpt-4o',
+      model: model ?? 'gpt-5.4',
       includeFuzzy: includeFuzzy === true,
     });
     const poContent = fs.readFileSync(poFilePath, 'utf8');
@@ -214,18 +207,6 @@ export function runTranslateCommand(args: TranslateCommandArgs): number | Promis
     }
   }
 
-  if (args.model != null) {
-    try {
-      validateModel(args.model);
-    } catch {
-      debugLogger.log('cli.runTranslateCommand', 'Model validation failed', {
-        model: args.model,
-      });
-      console.warn(getInvalidModelMessage(args.model));
-      return 1;
-    }
-  }
-
   if (!args.dryRun) {
     let resultApiKey: string;
     try {
diff --git a/src/translate.ts b/src/translate.ts
index bdd36cf..1c688ec 100644
--- a/src/translate.ts
+++ b/src/translate.ts
@@ -20,44 +20,7 @@ export type TranslateOptions = {
   context?: string;
 };
 
-const DEFAULT_MODEL = 'gpt-4o';
-const SUPPORTED_STRUCTURED_OUTPUT_MODEL_PATTERNS = [
-  /^gpt-4o(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-4o-mini(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-4\.1(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-4\.1-mini(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-4\.1-nano(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-5(?:-\d{4}-\d{2}-\d{2}|-chat-latest)?$/,
-  /^gpt-5-mini(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-5-nano(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-5-pro(?:-\d{4}-\d{2}-\d{2})?$/,
-  /^gpt-5\.1(?:-\d{4}-\d{2}-\d{2}|-chat-latest)?$/,
-  /^gpt-5\.2(?:-\d{4}-\d{2}-\d{2}|-chat-latest)?$/,
-  /^gpt-5-codex$/,
-  /^gpt-5\.1-codex$/,
-  /^gpt-5\.1-codex-mini$/,
-  /^gpt-5\.1-codex-max$/,
-  /^gpt-5\.2-codex$/,
-] as const;
-export const SUPPORTED_STRUCTURED_OUTPUT_MODELS = [
-  'gpt-4o',
-  'gpt-4o-mini',
-  'gpt-4.1',
-  'gpt-4.1-mini',
-  'gpt-4.1-nano',
-  'gpt-5',
-  'gpt-5-mini',
-  'gpt-5-nano',
-  'gpt-5-pro',
-  'gpt-5.1',
-  'gpt-5.2',
-  'gpt-5-codex',
-  'gpt-5.1-codex',
-  'gpt-5.1-codex-mini',
-  'gpt-5.1-codex-max',
-  'gpt-5.2-codex',
-] as const;
-
+const DEFAULT_MODEL = 'gpt-5.4';
 /** Error codes: https://developers.openai.com/api/docs/guides/error-codes#api-errors */
 
 const MAX_RETRIES = 3;
@@ -80,21 +43,6 @@ function isRetryableStatus(status: number): boolean {
   return status === 429 || status === 500 || status === 503;
 }
 
-function isSupportedStructuredOutputModel(model: string): boolean {
-  return SUPPORTED_STRUCTURED_OUTPUT_MODEL_PATTERNS.some((pattern) => pattern.test(model));
-}
-
-function validateStructuredOutputModel(model: string): void {
-  if (isSupportedStructuredOutputModel(model)) return;
-  throw new Error(
-    `Model "${model}" is not supported. This package requires an OpenAI Chat Completions model with json_schema structured outputs. Supported model families: ${SUPPORTED_STRUCTURED_OUTPUT_MODELS.join(', ')}.`,
-  );
-}
-
-export function validateModel(model: string): void {
-  validateStructuredOutputModel(model);
-}
-
 /** Request entry: either singular (msgid) or plural (msgid_plural). Optional msgctxt for gettext context. */
 export type TranslateRequestEntry =
   | { msgid: string; msgctxt?: string }
@@ -373,7 +321,6 @@ export async function translatePayload(
       apiKey: options.apiKey,
     });
   const model = options?.model ?? DEFAULT_MODEL;
-  validateStructuredOutputModel(model);
   debug.log('translate', 'Prepared translatePayload request summary', {
     model,
     target_language: payload.target_language,
diff --git a/test/cli/api-error-messages.test.ts b/test/cli/api-error-messages.test.ts
index e7b62f1..fa1937b 100644
--- a/test/cli/api-error-messages.test.ts
+++ b/test/cli/api-error-messages.test.ts
@@ -156,6 +156,46 @@ msgstr ""
   }
 });
 
+test('runTranslate shows structured output hint when 400 mentions json_schema', async () => {
+  mockState.errorToThrow = apiError(
+    400,
+    undefined,
+    "response_format of type 'json_schema' is not supported with this model",
+  );
+  const tempPo = getTmpPo(`
+msgid "Hello"
+msgstr ""
+`);
+
+  try {
+    const code = await runTranslate(tempPo.poFilePath, 'fake-key');
+    expect(code).toBe(1);
+    expect(consoleWarnSpy).toHaveBeenCalledWith(
+      expect.stringContaining('may not support json_schema structured outputs'),
+    );
+    expect(consoleWarnSpy.mock.calls[0][0]).toContain('gpt-5.4');
+  } finally {
+    tempPo.cleanup();
+  }
+});
+
+test('runTranslate shows generic 400 message for unrelated bad request', async () => {
+  mockState.errorToThrow = apiError(400, undefined, 'Invalid value for temperature');
+  const tempPo = getTmpPo(`
+msgid "Hello"
+msgstr ""
+`);
+
+  try {
+    const code = await runTranslate(tempPo.poFilePath, 'fake-key');
+    expect(code).toBe(1);
+    expect(consoleWarnSpy).toHaveBeenCalledWith(expect.stringContaining('Invalid request'));
+    expect(consoleWarnSpy.mock.calls[0][0]).toContain('Invalid value for temperature');
+  } finally {
+    tempPo.cleanup();
+  }
+});
+
 test('runTranslate shows generic message for non-API error', async () => {
   mockState.errorToThrow = new Error('Network connection failed');
   const tempPo = getTmpPo(`
diff --git a/test/cli/model.test.ts b/test/cli/model.test.ts
index 1470988..d7fc9e8 100644
--- a/test/cli/model.test.ts
+++ b/test/cli/model.test.ts
@@ -1,6 +1,4 @@
 import { test, expect, jest, beforeEach } from '@jest/globals';
-import path from 'node:path';
-import { spawnSync } from 'node:child_process';
 import { getTmpPo } from '../test-utils/getTmpPo';
 import { runTranslate } from '../../src/cli/runTranslate';
 import { translateStrings } from '../../src/translate';
@@ -19,26 +17,6 @@ beforeEach(() => {
   translateStringsMock.mockReset();
 });
 
-test('CLI exits with error when --model is invalid', () => {
-  const cliPath = path.resolve(process.cwd(), 'dist/src/cli/index.js');
-  const tempPo = getTmpPo(`
-msgid "Hello"
-msgstr ""
-`);
-
-  const runResult = spawnSync(
-    process.execPath,
-    [cliPath, tempPo.poFilePath, '--dry-run', '--model=not-a-real-model'],
-    { encoding: 'utf8' },
-  );
-
-  tempPo.cleanup();
-
-  expect(runResult.status).toBe(1);
-  expect(runResult.stderr).toContain('Invalid --model "not-a-real-model"');
-  expect(runResult.stderr).toMatch(/Supported model families/i);
-});
-
 test('runTranslate forwards model to translateStrings', async () => {
   translateStringsMock.mockResolvedValue([{ msgid: 'Hello', msgstr: 'Привіт' }]);
   const tempPo = getTmpPo(`
diff --git a/test/translate.test.ts b/test/translate.test.ts
index 58a43c8..5267e00 100644
--- a/test/translate.test.ts
+++ b/test/translate.test.ts
@@ -249,66 +249,6 @@ test('translatePayload throws when response missing translations', async () => {
   ).rejects.toThrow(/translations/i);
 });
 
-test('translatePayload rejects unsupported models before calling OpenAI', async () => {
-  const createMock = jest.fn<(params: unknown) => Promise<unknown>>();
-  const mockClient = { chat: { completions: { create: createMock } } } as unknown as OpenAI;
-
-  await expect(
-    translatePayload(
-      {
-        formula: '',
-        target_language: 'uk',
-        source_language: 'en',
-        translations: [{ msgid: 'Hi' }],
-      },
-      { apiKey: 'test-key', client: mockClient, model: 'gpt-4-turbo' },
-    ),
-  ).rejects.toThrow(/json_schema structured outputs|supported models/i);
-
-  expect(createMock).not.toHaveBeenCalled();
-});
-
-test('translatePayload allows supported GPT-5 structured-output models', async () => {
-  const createMock = jest
-    .fn<(params: unknown) => Promise<unknown>>()
-    .mockResolvedValue(
-      mockCompletion(JSON.stringify({ translations: [{ msgid: 'Hi', msgstr: 'Привіт' }] })),
-    );
-  const mockClient = { chat: { completions: { create: createMock } } } as unknown as OpenAI;
-
-  const result = await translatePayload(
-    {
-      formula: '',
-      target_language: 'uk',
-      source_language: 'en',
-      translations: [{ msgid: 'Hi' }],
-    },
-    { apiKey: 'test-key', client: mockClient, model: 'gpt-5.2' },
-  );
-
-  expect(result.translations[0]?.msgstr).toBe('Привіт');
-  expect(createMock).toHaveBeenCalledTimes(1);
-});
-
-test('translatePayload rejects GPT-5.2 models without structured outputs support', async () => {
-  const createMock = jest.fn<(params: unknown) => Promise<unknown>>();
-  const mockClient = { chat: { completions: { create: createMock } } } as unknown as OpenAI;
-
-  await expect(
-    translatePayload(
-      {
-        formula: '',
-        target_language: 'uk',
-        source_language: 'en',
-        translations: [{ msgid: 'Hi' }],
-      },
-      { apiKey: 'test-key', client: mockClient, model: 'gpt-5.2-pro' },
-    ),
-  ).rejects.toThrow(/not supported/i);
-
-  expect(createMock).not.toHaveBeenCalled();
-});
-
 test('translateItems sends items and returns translated strings in same order', async () => {
   const responsePayload = {
     translations: [