diff --git a/src/api/providers/__tests__/lmstudio.spec.ts b/src/api/providers/__tests__/lmstudio.spec.ts
index 0adebdeea7ac..bfecf3403662 100644
--- a/src/api/providers/__tests__/lmstudio.spec.ts
+++ b/src/api/providers/__tests__/lmstudio.spec.ts
@@ -114,8 +114,9 @@ describe("LmStudioHandler", () => {
 			expect(textChunks[0].text).toBe("Test response")
 		})
 
-		it("should handle API errors", async () => {
-			mockCreate.mockRejectedValueOnce(new Error("API Error"))
+		it("should handle connection errors", async () => {
+			const connectionError = new Error("connect ECONNREFUSED 127.0.0.1:1234")
+			mockCreate.mockRejectedValueOnce(connectionError)
 
 			const stream = handler.createMessage(systemPrompt, messages)
 
@@ -123,7 +124,45 @@ describe("LmStudioHandler", () => {
 				for await (const _chunk of stream) {
 					// Should not reach here
 				}
-			}).rejects.toThrow("Please check the LM Studio developer logs to debug what went wrong")
+			}).rejects.toThrow("Cannot connect to LM Studio at http://localhost:1234")
+		})
+
+		it("should handle model not found errors", async () => {
+			const modelError = new Error("model 'local-model' not found")
+			mockCreate.mockRejectedValueOnce(modelError)
+
+			const stream = handler.createMessage(systemPrompt, messages)
+
+			await expect(async () => {
+				for await (const _chunk of stream) {
+					// Should not reach here
+				}
+			}).rejects.toThrow('Model "local-model" not found in LM Studio')
+		})
+
+		it("should handle context length errors", async () => {
+			const contextError = new Error("context length exceeded")
+			mockCreate.mockRejectedValueOnce(contextError)
+
+			const stream = handler.createMessage(systemPrompt, messages)
+
+			await expect(async () => {
+				for await (const _chunk of stream) {
+					// Should not reach here
+				}
+			}).rejects.toThrow("Context length exceeded")
+		})
+
+		it("should handle generic API errors", async () => {
+			mockCreate.mockRejectedValueOnce(new Error("Unknown API Error"))
+
+			const stream = handler.createMessage(systemPrompt, messages)
+
+			await expect(async () => {
+				for await (const _chunk of stream) {
+					// Should not reach here
+				}
+			}).rejects.toThrow("LM Studio completion error")
 		})
 	})
 
@@ -139,13 +178,33 @@ describe("LmStudioHandler", () => {
 			})
 		})
 
-		it("should handle API errors", async () => {
-			mockCreate.mockRejectedValueOnce(new Error("API Error"))
+		it("should handle connection errors", async () => {
+			const connectionError = new Error("connect ECONNREFUSED 127.0.0.1:1234")
+			mockCreate.mockRejectedValueOnce(connectionError)
 			await expect(handler.completePrompt("Test prompt")).rejects.toThrow(
-				"Please check the LM Studio developer logs to debug what went wrong",
+				"Cannot connect to LM Studio at http://localhost:1234",
 			)
 		})
 
+		it("should handle model not found errors", async () => {
+			const modelError = new Error("model 'local-model' not found")
+			mockCreate.mockRejectedValueOnce(modelError)
+			await expect(handler.completePrompt("Test prompt")).rejects.toThrow(
+				'Model "local-model" not found in LM Studio',
+			)
+		})
+
+		it("should handle context length errors", async () => {
+			const contextError = new Error("token limit exceeded")
+			mockCreate.mockRejectedValueOnce(contextError)
+			await expect(handler.completePrompt("Test prompt")).rejects.toThrow("Context length exceeded")
+		})
+
+		it("should handle generic API errors", async () => {
+			mockCreate.mockRejectedValueOnce(new Error("Unknown API Error"))
+			await expect(handler.completePrompt("Test prompt")).rejects.toThrow("LM Studio completion error")
+		})
+
 		it("should handle empty response", async () => {
 			mockCreate.mockResolvedValueOnce({
 				choices: [{ message: { content: "" } }],
diff --git a/src/api/providers/lm-studio.ts b/src/api/providers/lm-studio.ts
index 6c58a96ae1fa..de7aaa96dbd1 100644
--- a/src/api/providers/lm-studio.ts
+++ b/src/api/providers/lm-studio.ts
@@ -97,6 +97,38 @@ export class LmStudioHandler extends BaseProvider implements SingleCompletionHan
 			try {
 				results = await this.client.chat.completions.create(params)
 			} catch (error) {
+				// Handle specific error cases
+				const errorMessage = error instanceof Error ? error.message : String(error)
+
+				// Check for connection errors
+				if (errorMessage.includes("ECONNREFUSED") || errorMessage.includes("ENOTFOUND")) {
+					throw new Error(
+						`Cannot connect to LM Studio at ${this.options.lmStudioBaseUrl || "http://localhost:1234"}. Please ensure LM Studio is running and the server is started.`,
+					)
+				}
+
+				// Check for model not found errors
+				if (
+					errorMessage.includes("model") &&
+					(errorMessage.includes("not found") || errorMessage.includes("does not exist"))
+				) {
+					throw new Error(
+						`Model "${this.getModel().id}" not found in LM Studio. Please ensure the model is loaded in LM Studio.`,
+					)
+				}
+
+				// Check for context length errors
+				if (
+					errorMessage.includes("context") ||
+					errorMessage.includes("token") ||
+					errorMessage.includes("length")
+				) {
+					throw new Error(
+						`Context length exceeded for model "${this.getModel().id}". Please load the model with a larger context window in LM Studio, or use a different model that supports longer contexts.`,
+					)
+				}
+
+				// Use the enhanced error handler for other OpenAI-like errors
 				throw handleOpenAIError(error, this.providerName)
 			}
 
@@ -138,8 +170,14 @@ export class LmStudioHandler extends BaseProvider implements SingleCompletionHan
 				outputTokens,
 			} as const
 		} catch (error) {
+			// If error was already processed and re-thrown above, just re-throw it
+			if (error instanceof Error && error.message.includes("LM Studio")) {
+				throw error
+			}
+
+			// Generic fallback error
 			throw new Error(
-				"Please check the LM Studio developer logs to debug what went wrong. You may need to load the model with a larger context length to work with Roo Code's prompts.",
+				`LM Studio error: ${error instanceof Error ? error.message : String(error)}. Please check the LM Studio developer logs for more details.`,
 			)
 		}
 	}
@@ -178,12 +216,50 @@ export class LmStudioHandler extends BaseProvider implements SingleCompletionHan
 			try {
 				response = await this.client.chat.completions.create(params)
 			} catch (error) {
+				// Handle specific error cases
+				const errorMessage = error instanceof Error ? error.message : String(error)
+
+				// Check for connection errors
+				if (errorMessage.includes("ECONNREFUSED") || errorMessage.includes("ENOTFOUND")) {
+					throw new Error(
+						`Cannot connect to LM Studio at ${this.options.lmStudioBaseUrl || "http://localhost:1234"}. Please ensure LM Studio is running and the server is started.`,
+					)
+				}
+
+				// Check for model not found errors
+				if (
+					errorMessage.includes("model") &&
+					(errorMessage.includes("not found") || errorMessage.includes("does not exist"))
+				) {
+					throw new Error(
+						`Model "${this.getModel().id}" not found in LM Studio. Please ensure the model is loaded in LM Studio.`,
+					)
+				}
+
+				// Check for context length errors
+				if (
+					errorMessage.includes("context") ||
+					errorMessage.includes("token") ||
+					errorMessage.includes("length")
+				) {
+					throw new Error(
+						`Context length exceeded for model "${this.getModel().id}". Please load the model with a larger context window in LM Studio, or use a different model that supports longer contexts.`,
+					)
+				}
+
+				// Use the enhanced error handler for other OpenAI-like errors
 				throw handleOpenAIError(error, this.providerName)
 			}
 			return response.choices[0]?.message.content || ""
 		} catch (error) {
+			// If error was already processed and re-thrown above, just re-throw it
+			if (error instanceof Error && error.message.includes("LM Studio")) {
+				throw error
+			}
+
+			// Generic fallback error
 			throw new Error(
-				"Please check the LM Studio developer logs to debug what went wrong. You may need to load the model with a larger context length to work with Roo Code's prompts.",
+				`LM Studio error: ${error instanceof Error ? error.message : String(error)}. Please check the LM Studio developer logs for more details.`,
 			)
 		}
 	}