@picsart/ai-sdk 5.8.0 → 5.10.0

package/index.d.ts

@@ -251,6 +251,11 @@ type ModelInputById = {
         imageUrls?: string[];
         thinking?: "off" | "low" | "medium" | "high";
     };
+    "gemini-3.7-flash": {
+        prompt: string;
+        imageUrls?: string[];
+        thinking?: "off" | "low" | "medium" | "high";
+    };
     "gemini-omni-flash-preview": {
         prompt: string;
         aspectRatio?: "16:9" | "9:16";
@@ -305,6 +310,14 @@ type ModelInputById = {
         count?: 1 | 2 | 4;
         imageUrls?: string[];
     };
+    "grok-imagine-image-2.0": {
+        prompt: string;
+        aspectRatio?: "1:1" | "16:9" | "9:16" | "4:3" | "3:4" | "3:2" | "2:3" | "2:1" | "1:2" | "19.5:9" | "9:19.5" | "20:9" | "9:20";
+        resolution?: "1k" | "2k";
+        quality?: "low" | "medium";
+        count?: 1 | 2 | 4;
+        imageUrls?: string[];
+    };
     "grok-imagine-image-quality": {
         prompt: string;
         aspectRatio?: "1:1" | "16:9" | "9:16" | "4:3" | "3:4" | "3:2" | "2:3" | "2:1" | "1:2" | "19.5:9" | "9:19.5" | "20:9" | "9:20";
@@ -1515,7 +1528,7 @@ type ModelInputById = {
 type TypedModelId = keyof ModelInputById;
 type ModelInput<M extends TypedModelId> = ModelInputById[M];
 /** IDs of text-generation (LLM) models — narrows generateText(). */
-type TextModelId = "claude-haiku-4-5" | "claude-opus-4-8" | "claude-sonnet-4-6" | "gemini-3-pro" | "gemini-3.5-flash-lite" | "gemini-3.6-flash" | "gpt-5.5";
+type TextModelId = "claude-haiku-4-5" | "claude-opus-4-8" | "claude-sonnet-4-6" | "gemini-3-pro" | "gemini-3.5-flash-lite" | "gemini-3.6-flash" | "gemini-3.7-flash" | "gpt-5.5";
 type TextModelInputById = Pick<ModelInputById, TextModelId>;
 
 /**
@@ -2458,6 +2471,7 @@ declare const Models: {
     readonly Gemini31FlashLiteImage: "gemini-3.1-flash-lite-image";
     readonly Gemini35FlashLite: "gemini-3.5-flash-lite";
     readonly Gemini36Flash: "gemini-3.6-flash";
+    readonly Gemini37Flash: "gemini-3.7-flash";
     readonly GeminiOmniFlashPreview: "gemini-omni-flash-preview";
     readonly Gpt55: "gpt-5.5";
     readonly GptImage1: "gpt-image-1";
@@ -2466,6 +2480,7 @@ declare const Models: {
     readonly GrokEditVideo: "grok-edit-video";
     readonly GrokExtendVideo: "grok-extend-video";
     readonly GrokImagineImage: "grok-imagine-image";
+    readonly GrokImagineImage20: "grok-imagine-image-2.0";
     readonly GrokImagineImageQuality: "grok-imagine-image-quality";
     readonly GrokImagineVideo: "grok-imagine-video";
     readonly GrokImagineVideo15: "grok-imagine-video-1.5";

package/index.js

@@ -4890,6 +4890,28 @@ var { MODELS: MODELS17 } = defineModels("grok", [
       ...params.imageInput(1, "Source Image")
     }
   },
+  {
+    id: "grok-imagine-image-2.0",
+    name: "Grok Imagine 2.0",
+    addedAt: "2026-08-19",
+    workflow: "x-ai/v1/images/generations",
+    editWorkflow: "x-ai/v1/images/edits",
+    estimatedTime: 16,
+    mode: "image",
+    inputType: "t2i",
+    description: "Latest Grok Imagine generation \u2014 sharper detail with a low/medium quality tier.",
+    features: [feat("Image Input", "input"), feat("2k", "resolution")],
+    paramConfig: {
+      ...params.prompt(),
+      ...params.aspectRatio(GROK_IMAGE_AR, "1:1"),
+      ...params.resolution(GROK_IMAGE_RESOLUTIONS, "1k"),
+      // Vendor-side default is medium; only supported by grok-imagine-image-2.0
+      // (generations only — the edits command has no quality field).
+      ...p.quality(["low", "medium"], "medium"),
+      ...params.count([1, 2, 4]),
+      ...params.imageInput(1, "Source Image")
+    }
+  },
   // ── Audio ─────────────────────────────────────────
   {
     id: "grok-tts",
@@ -4911,6 +4933,33 @@ var { MODELS: MODELS17 } = defineModels("grok", [
   }
 ]);
 
+// src/vendors/catalog/grok.payloads.ts
+var buildGrokImage2Payload = (input) => ({
+  model: "grok-imagine-image-2.0",
+  prompt: input.prompt,
+  n: input.count ?? 1,
+  ...input.aspectRatio ? { aspect_ratio: input.aspectRatio } : {},
+  ...input.resolution ? { resolution: input.resolution } : {},
+  ...input.quality ? { quality: input.quality } : {}
+});
+var buildGrokImage2EditPayload = (input) => {
+  const urls = input.imageUrls ?? [];
+  const imagePart2 = urls.length > 1 ? { images: urls.map((url) => ({ url })) } : urls.length === 1 ? { image: { url: urls[0] } } : {};
+  return {
+    model: "grok-imagine-image-2.0",
+    prompt: input.prompt,
+    n: input.count ?? 1,
+    ...input.resolution ? { resolution: input.resolution } : {},
+    ...imagePart2
+  };
+};
+registerPayloads(MODELS17, {
+  "grok-imagine-image-2.0": buildGrokImage2Payload
+});
+registerEditPayloads(MODELS17, {
+  "grok-imagine-image-2.0": buildGrokImage2EditPayload
+});
+
 // src/vendors/catalog/pika.ts
 var PIKA_RATIO_MAP = {
   "16:9": 1.77,
@@ -8215,6 +8264,23 @@ var { MODELS: GEMINI_LLM } = defineModels("google", [
       ...thinkingParam(["low", "high"])
     }
   },
+  {
+    id: "gemini-3.7-flash",
+    name: "Gemini 3.7 Flash",
+    workflow: "chat-completions",
+    addedAt: "2026-08-19",
+    estimatedTime: 5,
+    mode: "text",
+    inputType: "i2t",
+    badge: ["fast"],
+    description: "Latest fast Gemini model \u2014 low-latency multimodal text generation.",
+    features: [feat("Vision", "input"), feat("Thinking", "characteristic")],
+    paramConfig: {
+      ...params.prompt(),
+      ...params.imageInput(8, "Images"),
+      ...thinkingParam(["low", "medium", "high"])
+    }
+  },
   {
     id: "gemini-3.6-flash",
     name: "Gemini 3.6 Flash",
@@ -8308,6 +8374,7 @@ registerPayloads(MODELS36, {
   "gemini-3-pro": buildGeminiPayload("gemini-3-pro-preview"),
   // Flash models route through chat-completions (OpenAI-shaped), not the
   // native `gemini` workflow. flash-lite has no thinking param → reasoning_effort omitted.
+  "gemini-3.7-flash": buildOpenAiPayload("gemini-3.7-flash"),
   "gemini-3.6-flash": buildOpenAiPayload("gemini-3.6-flash"),
   "gemini-3.5-flash-lite": buildOpenAiPayload("gemini-3.5-flash-lite")
 });
@@ -10447,6 +10514,7 @@ var Gemini31FlashImage = "gemini-3.1-flash-image";
 var Gemini31FlashLiteImage = "gemini-3.1-flash-lite-image";
 var Gemini35FlashLite = "gemini-3.5-flash-lite";
 var Gemini36Flash = "gemini-3.6-flash";
+var Gemini37Flash = "gemini-3.7-flash";
 var GeminiOmniFlashPreview = "gemini-omni-flash-preview";
 var Gpt55 = "gpt-5.5";
 var GptImage1 = "gpt-image-1";
@@ -10455,6 +10523,7 @@ var GptImage2 = "gpt-image-2";
 var GrokEditVideo = "grok-edit-video";
 var GrokExtendVideo = "grok-extend-video";
 var GrokImagineImage = "grok-imagine-image";
+var GrokImagineImage20 = "grok-imagine-image-2.0";
 var GrokImagineImageQuality = "grok-imagine-image-quality";
 var GrokImagineVideo = "grok-imagine-video";
 var GrokImagineVideo15 = "grok-imagine-video-1.5";
@@ -10655,6 +10724,7 @@ var Models = {
   Gemini31FlashLiteImage,
   Gemini35FlashLite,
   Gemini36Flash,
+  Gemini37Flash,
   GeminiOmniFlashPreview,
   Gpt55,
   GptImage1,
@@ -10663,6 +10733,7 @@ var Models = {
   GrokEditVideo,
   GrokExtendVideo,
   GrokImagineImage,
+  GrokImagineImage20,
   GrokImagineImageQuality,
   GrokImagineVideo,
   GrokImagineVideo15,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@picsart/ai-sdk",
-  "version": "5.8.0",
+  "version": "5.10.0",
   "type": "module",
   "description": "Type-safe SDK for 100+ AI models — image, video, audio, and text generation with Picsart",
   "license": "MIT",