fix(providers): always use optional instead of mix of nullish for providerOptions (vercel#6122)

## Background

We were using nullish in some places (due to AI autocompletion)

## Summary

Now it's unified to use `.optional()` everywhere

## Verification

Checked JSDoc to make sure we're not breaking anything

Author: samdenty

.changeset/weak-moles-nail.md

@@ -0,0 +1,9 @@
+---
+'@ai-sdk/amazon-bedrock': patch
+'@ai-sdk/mistral': patch
+'@ai-sdk/cohere': patch
+'@ai-sdk/openai': patch
+'@ai-sdk/groq': patch
+---
+
+fix(providers): always use optional instead of mix of nullish for providerOptions

packages/amazon-bedrock/src/bedrock-chat-options.ts

@@ -44,10 +44,10 @@ export const bedrockProviderOptions = z.object({
   additionalModelRequestFields: z.record(z.any()).optional(),
   reasoningConfig: z
     .object({
-      type: z.union([z.literal('enabled'), z.literal('disabled')]).nullish(),
-      budgetTokens: z.number().nullish(),
+      type: z.union([z.literal('enabled'), z.literal('disabled')]).optional(),
+      budgetTokens: z.number().optional(),
     })
-    .nullish(),
+    .optional(),
 });
 
 export type BedrockProviderOptions = z.infer<typeof bedrockProviderOptions>;

packages/cohere/src/cohere-embedding-options.ts

@@ -21,7 +21,7 @@ export const cohereEmbeddingOptions = z.object({
    */
   inputType: z
     .enum(['search_document', 'search_query', 'classification', 'clustering'])
-    .nullish(),
+    .optional(),
 
   /**
    * Specifies how the API will handle inputs longer than the maximum token length.
@@ -31,7 +31,7 @@ export const cohereEmbeddingOptions = z.object({
    * - "START": Will discard the start of the input until the remaining input is exactly the maximum input token length for the model.
    * - "END": Will discard the end of the input until the remaining input is exactly the maximum input token length for the model.
    */
-  truncate: z.enum(['NONE', 'START', 'END']).nullish(),
+  truncate: z.enum(['NONE', 'START', 'END']).optional(),
 });
 
 export type CohereEmbeddingOptions = z.infer<typeof cohereEmbeddingOptions>;

packages/groq/src/groq-chat-options.ts

@@ -20,18 +20,18 @@ export type GroqChatModelId =
   | (string & {});
 
 export const groqProviderOptions = z.object({
-  reasoningFormat: z.enum(['parsed', 'raw', 'hidden']).nullish(),
+  reasoningFormat: z.enum(['parsed', 'raw', 'hidden']).optional(),
 
   /**
    * Whether to enable parallel function calling during tool use. Default to true.
    */
-  parallelToolCalls: z.boolean().nullish(),
+  parallelToolCalls: z.boolean().optional(),
 
   /**
    * A unique identifier representing your end-user, which can help OpenAI to
    * monitor and detect abuse. Learn more.
    */
-  user: z.string().nullish(),
+  user: z.string().optional(),
 });
 
 export type GroqProviderOptions = z.infer<typeof groqProviderOptions>;

packages/mistral/src/mistral-chat-options.ts

@@ -22,10 +22,10 @@ Whether to inject a safety prompt before all conversations.
 
 Defaults to `false`.
    */
-  safePrompt: z.boolean().nullish(),
+  safePrompt: z.boolean().optional(),
 
-  documentImageLimit: z.number().nullish(),
-  documentPageLimit: z.number().nullish(),
+  documentImageLimit: z.number().optional(),
+  documentPageLimit: z.number().optional(),
 });
 
 export type MistralProviderOptions = z.infer<typeof mistralProviderOptions>;

packages/openai/src/openai-transcription-options.ts

@@ -12,23 +12,23 @@ export const openAITranscriptionProviderOptions = z.object({
    * Additional information to include in the transcription response.
    */
 
-  include: z.array(z.string()).nullish(),
+  include: z.array(z.string()).optional(),
 
   /**
    * The language of the input audio in ISO-639-1 format.
    */
-  language: z.string().nullish(),
+  language: z.string().optional(),
 
   /**
    * An optional text to guide the model's style or continue a previous audio segment.
    */
-  prompt: z.string().nullish(),
+  prompt: z.string().optional(),
 
   /**
    * The sampling temperature, between 0 and 1.
    * @default 0
    */
-  temperature: z.number().min(0).max(1).default(0).nullish(),
+  temperature: z.number().min(0).max(1).default(0).optional(),
 
   /**
    * The timestamp granularities to populate for this transcription.
@@ -37,7 +37,7 @@ export const openAITranscriptionProviderOptions = z.object({
   timestampGranularities: z
     .array(z.enum(['word', 'segment']))
     .default(['segment'])
-    .nullish(),
+    .optional(),
 });
 
 export type OpenAITranscriptionProviderOptions = z.infer<