diff --git a/packages/ai/src/protocols/openai-images.ts b/packages/ai/src/protocols/openai-images.ts
index 6ff18e775705..f059d9aa991a 100644
--- a/packages/ai/src/protocols/openai-images.ts
+++ b/packages/ai/src/protocols/openai-images.ts
@@ -13,7 +13,7 @@ import { Framing } from "../route/framing.js"
 import { MediaProtocol } from "../route/media-protocol.js"
 import { MediaRoute } from "../route/media.js"
 import { AIError, mergeJsonRecords, ProviderID, tokenUsage, type MediaUsage, type OpenString } from "../schema/index.js"
-import { optionalNull, ProviderShared } from "./shared.js"
+import { optionalNullish, optionalNull, ProviderShared } from "./shared.js"
 import { ImageSizing } from "./utils/image-sizing.js"
 import { MediaInput } from "./utils/media-input.js"
 
@@ -56,18 +56,21 @@ const Usage = Schema.Struct({
 })
 
 /** What the provider actually rendered; it can differ from the request when `auto` or a default applied. */
+// `optionalNullish` (rather than `Schema.optional`) tolerates OpenAI-compatible hosts that report an absent optional
+// field (e.g. `quality`, `background`, or `usage`) as an explicit `null` instead of omitting it, and normalizes it to
+// `undefined`. This matches the null-tolerant fields already used for `data[].url` / `data[].b64_json`.
 const Settings = {
-  output_format: Schema.optional(Schema.String),
-  size: Schema.optional(Schema.String),
-  quality: Schema.optional(Schema.String),
-  background: Schema.optional(Schema.String),
+  output_format: optionalNullish(Schema.String),
+  size: optionalNullish(Schema.String),
+  quality: optionalNullish(Schema.String),
+  background: optionalNullish(Schema.String),
 }
 
 const OpenAIImageResponse = Schema.Struct({
   // GPT Image models always return `b64_json`; compatible hosts that default to `url` can return either.
   data: Schema.Array(Schema.Struct({ b64_json: optionalNull(Schema.String), url: optionalNull(Schema.String) })),
   ...Settings,
-  usage: Schema.optional(Usage),
+  usage: optionalNullish(Usage),
 })
 
 // ---------------------------------------------------------------------------
diff --git a/packages/ai/test/image.test.ts b/packages/ai/test/image.test.ts
index 1a8f1f5aca44..b76b0d6504aa 100644
--- a/packages/ai/test/image.test.ts
+++ b/packages/ai/test/image.test.ts
@@ -239,6 +239,43 @@ describe("Image", () => {
     )
   })
 
+  it.effect(
+    "decodes a response whose optional fields are null instead of omitted (OpenAI-compatible hosts)",
+    () =>
+      Effect.gen(function* () {
+        const response = yield* ai.image.generate({
+          model: OpenAI.configure({ apiKey: "test", baseURL: "https://openai.test/v1" }).image("gpt-image-2"),
+          prompt: "A lighthouse",
+        }).pipe(
+          Effect.provide(
+            layer((input) =>
+              Effect.succeed(
+                json(input, {
+                  background: null,
+                  data: [{ b64_json: "AQID", url: null }],
+                  output_format: "png",
+                  quality: null,
+                  size: "1024x1024",
+                  usage: null,
+                }),
+              ),
+            ),
+          ),
+        )
+        expect(response.images).toHaveLength(1)
+        expect(response.image.source).toEqual({
+          type: "bytes",
+          data: Uint8Array.from([1, 2, 3]),
+          mediaType: "image/png",
+        })
+        expect(response.image.info).toEqual({ format: "png", width: 1024, height: 1024 })
+        expect(response.usage).toBeUndefined()
+        expect(response.providerMetadata).toEqual({
+          openai: { outputFormat: "png", size: "1024x1024" },
+        })
+      }),
+  )
+
   it.effect("classifies an OpenAI stream moderation error event as a content policy rejection", () =>
     Effect.gen(function* () {
       const error = yield* ai.image