feat(zai): add GLM-4.5V vision model (supportsImages, pricing, 16K max output); add tests

Author: hannesrudolph

packages/types/src/providers/zai.ts

@@ -66,6 +66,18 @@ export const internationalZAiModels = {
 		cacheReadsPrice: 0,
 		description: "GLM-4.5-Flash is a free, high-speed model excellent for reasoning, coding, and agentic tasks.",
 	},
+	"glm-4.5v": {
+		maxTokens: 16_384,
+		contextWindow: 131_072,
+		supportsImages: true,
+		supportsPromptCache: true,
+		inputPrice: 0.6,
+		outputPrice: 1.8,
+		cacheWritesPrice: 0,
+		cacheReadsPrice: 0.11,
+		description:
+			"GLM-4.5V is Z.AI's multimodal visual reasoning model (image/video/text/file input), optimized for GUI tasks, grounding, and document/video understanding.",
+	},
 	"glm-4.6": {
 		maxTokens: 98_304,
 		contextWindow: 204_800,
@@ -181,6 +193,18 @@ export const mainlandZAiModels = {
 		cacheReadsPrice: 0,
 		description: "GLM-4.5-Flash is a free, high-speed model excellent for reasoning, coding, and agentic tasks.",
 	},
+	"glm-4.5v": {
+		maxTokens: 16_384,
+		contextWindow: 131_072,
+		supportsImages: true,
+		supportsPromptCache: true,
+		inputPrice: 0.29,
+		outputPrice: 0.93,
+		cacheWritesPrice: 0,
+		cacheReadsPrice: 0.057,
+		description:
+			"GLM-4.5V is Z.AI's multimodal visual reasoning model (image/video/text/file input), optimized for GUI tasks, grounding, and document/video understanding.",
+	},
 	"glm-4.6": {
 		maxTokens: 98_304,
 		contextWindow: 204_800,

src/api/providers/__tests__/zai.spec.ts

@@ -84,6 +84,21 @@ describe("ZAiHandler", () => {
 			expect(model.info).toEqual(internationalZAiModels[testModelId])
 			expect(model.info.contextWindow).toBe(204_800)
 		})
+
+		it("should return GLM-4.5v international model with vision support", () => {
+			const testModelId: InternationalZAiModelId = "glm-4.5v"
+			const handlerWithModel = new ZAiHandler({
+				apiModelId: testModelId,
+				zaiApiKey: "test-zai-api-key",
+				zaiApiLine: "international_coding",
+			})
+			const model = handlerWithModel.getModel()
+			expect(model.id).toBe(testModelId)
+			expect(model.info).toEqual(internationalZAiModels[testModelId])
+			expect(model.info.supportsImages).toBe(true)
+			expect(model.info.maxTokens).toBe(16_384)
+			expect(model.info.contextWindow).toBe(131_072)
+		})
 	})
 
 	describe("China Z AI", () => {
@@ -134,6 +149,21 @@ describe("ZAiHandler", () => {
 			expect(model.info).toEqual(mainlandZAiModels[testModelId])
 			expect(model.info.contextWindow).toBe(204_800)
 		})
+
+		it("should return GLM-4.5v China model with vision support", () => {
+			const testModelId: MainlandZAiModelId = "glm-4.5v"
+			const handlerWithModel = new ZAiHandler({
+				apiModelId: testModelId,
+				zaiApiKey: "test-zai-api-key",
+				zaiApiLine: "china_coding",
+			})
+			const model = handlerWithModel.getModel()
+			expect(model.id).toBe(testModelId)
+			expect(model.info).toEqual(mainlandZAiModels[testModelId])
+			expect(model.info.supportsImages).toBe(true)
+			expect(model.info.maxTokens).toBe(16_384)
+			expect(model.info.contextWindow).toBe(131_072)
+		})
 	})
 
 	describe("Default behavior", () => {