@oh-my-pi/pi-ai 11.2.2 → 11.2.3

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@oh-my-pi/pi-ai",
-	"version": "11.2.2",
+	"version": "11.2.3",
 	"description": "Unified LLM API with automatic model discovery and provider configuration",
 	"type": "module",
 	"main": "./src/index.ts",
@@ -63,7 +63,7 @@
 		"@connectrpc/connect-node": "^2.1.1",
 		"@google/genai": "^1.39.0",
 		"@mistralai/mistralai": "^1.13.0",
-		"@oh-my-pi/pi-utils": "11.2.2",
+		"@oh-my-pi/pi-utils": "11.2.3",
 		"@sinclair/typebox": "^0.34.48",
 		"@smithy/node-http-handler": "^4.4.9",
 		"ajv": "^8.17.1",

package/src/models.generated.ts

@@ -107,6 +107,23 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 4096,
 		} satisfies Model<"bedrock-converse-stream">,
+		"anthropic.claude-opus-4-6-v1:0": {
+			id: "anthropic.claude-opus-4-6-v1:0",
+			name: "Claude Opus 4.6",
+			api: "bedrock-converse-stream",
+			provider: "amazon-bedrock",
+			baseUrl: "https://bedrock-runtime.us-east-1.amazonaws.com",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 0.5,
+				cacheWrite: 6.25,
+			},
+			contextWindow: 200000,
+			maxTokens: 128000,
+		} satisfies Model<"bedrock-converse-stream">,
 		"cohere.command-r-plus-v1:0": {
 			id: "cohere.command-r-plus-v1:0",
 			name: "Command R+",
@@ -192,6 +209,23 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 64000,
 		} satisfies Model<"bedrock-converse-stream">,
+		"eu.anthropic.claude-opus-4-6-v1:0": {
+			id: "eu.anthropic.claude-opus-4-6-v1:0",
+			name: "Claude Opus 4.6 (EU)",
+			api: "bedrock-converse-stream",
+			provider: "amazon-bedrock",
+			baseUrl: "https://bedrock-runtime.us-east-1.amazonaws.com",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 1.5,
+				cacheWrite: 18.75,
+			},
+			contextWindow: 200000,
+			maxTokens: 128000,
+		} satisfies Model<"bedrock-converse-stream">,
 		"eu.anthropic.claude-sonnet-4-20250514-v1:0": {
 			id: "eu.anthropic.claude-sonnet-4-20250514-v1:0",
 			name: "Claude Sonnet 4 (EU)",
@@ -277,6 +311,23 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 64000,
 		} satisfies Model<"bedrock-converse-stream">,
+		"global.anthropic.claude-opus-4-6-v1:0": {
+			id: "global.anthropic.claude-opus-4-6-v1:0",
+			name: "Claude Opus 4.6 (Global)",
+			api: "bedrock-converse-stream",
+			provider: "amazon-bedrock",
+			baseUrl: "https://bedrock-runtime.us-east-1.amazonaws.com",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 0.5,
+				cacheWrite: 6.25,
+			},
+			contextWindow: 200000,
+			maxTokens: 128000,
+		} satisfies Model<"bedrock-converse-stream">,
 		"global.anthropic.claude-sonnet-4-20250514-v1:0": {
 			id: "global.anthropic.claude-sonnet-4-20250514-v1:0",
 			name: "Claude Sonnet 4",
@@ -855,6 +906,23 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 64000,
 		} satisfies Model<"bedrock-converse-stream">,
+		"us.anthropic.claude-opus-4-6-v1:0": {
+			id: "us.anthropic.claude-opus-4-6-v1:0",
+			name: "Claude Opus 4.6 (US)",
+			api: "bedrock-converse-stream",
+			provider: "amazon-bedrock",
+			baseUrl: "https://bedrock-runtime.us-east-1.amazonaws.com",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 1.5,
+				cacheWrite: 18.75,
+			},
+			contextWindow: 200000,
+			maxTokens: 128000,
+		} satisfies Model<"bedrock-converse-stream">,
 		"us.anthropic.claude-sonnet-4-20250514-v1:0": {
 			id: "us.anthropic.claude-sonnet-4-20250514-v1:0",
 			name: "Claude Sonnet 4 (US)",
@@ -1316,6 +1384,40 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 64000,
 		} satisfies Model<"anthropic-messages">,
+		"claude-opus-4-6": {
+			id: "claude-opus-4-6",
+			name: "Claude Opus 4.6",
+			api: "anthropic-messages",
+			provider: "anthropic",
+			baseUrl: "https://api.anthropic.com",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 0.5,
+				cacheWrite: 6.25,
+			},
+			contextWindow: 1000000,
+			maxTokens: 128000,
+		} satisfies Model<"anthropic-messages">,
+		"claude-opus-4-6-20260205": {
+			id: "claude-opus-4-6-20260205",
+			name: "Claude Opus 4.6",
+			api: "anthropic-messages",
+			provider: "anthropic",
+			baseUrl: "https://api.anthropic.com",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 0.5,
+				cacheWrite: 6.25,
+			},
+			contextWindow: 200000,
+			maxTokens: 128000,
+		} satisfies Model<"anthropic-messages">,
 		"claude-sonnet-4-0": {
 			id: "claude-sonnet-4-0",
 			name: "Claude Sonnet 4 (latest)",
@@ -1700,6 +1802,25 @@ export const MODELS = {
 			contextWindow: 128000,
 			maxTokens: 16000,
 		} satisfies Model<"openai-completions">,
+		"claude-opus-4.6": {
+			id: "claude-opus-4.6",
+			name: "Claude Opus 4.6",
+			api: "openai-completions",
+			provider: "github-copilot",
+			baseUrl: "https://api.individual.githubcopilot.com",
+			headers: {"User-Agent":"GitHubCopilotChat/0.35.0","Editor-Version":"vscode/1.107.0","Editor-Plugin-Version":"copilot-chat/0.35.0","Copilot-Integration-Id":"vscode-chat"},
+			compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 0,
+				output: 0,
+				cacheRead: 0,
+				cacheWrite: 0,
+			},
+			contextWindow: 128000,
+			maxTokens: 16000,
+		} satisfies Model<"openai-completions">,
 		"claude-sonnet-4": {
 			id: "claude-sonnet-4",
 			name: "Claude Sonnet 4",
@@ -3030,63 +3151,6 @@ export const MODELS = {
 			contextWindow: 262144,
 			maxTokens: 32000,
 		} satisfies Model<"openai-completions">,
-		"kimi-k2": {
-			id: "kimi-k2",
-			name: "Kimi K2",
-			api: "openai-completions",
-			provider: "kimi-code",
-			baseUrl: "https://api.kimi.com/coding/v1",
-			headers: {"User-Agent":"KimiCLI/1.0","X-Msh-Platform":"kimi_cli"},
-			compat: {"thinkingFormat":"zai","reasoningContentField":"reasoning_content","supportsDeveloperRole":false},
-			reasoning: true,
-			input: ["text"],
-			cost: {
-				input: 0,
-				output: 0,
-				cacheRead: 0,
-				cacheWrite: 0,
-			},
-			contextWindow: 262144,
-			maxTokens: 32000,
-		} satisfies Model<"openai-completions">,
-		"kimi-k2-turbo-preview": {
-			id: "kimi-k2-turbo-preview",
-			name: "Kimi K2 Turbo Preview",
-			api: "openai-completions",
-			provider: "kimi-code",
-			baseUrl: "https://api.kimi.com/coding/v1",
-			headers: {"User-Agent":"KimiCLI/1.0","X-Msh-Platform":"kimi_cli"},
-			compat: {"thinkingFormat":"zai","reasoningContentField":"reasoning_content","supportsDeveloperRole":false},
-			reasoning: true,
-			input: ["text"],
-			cost: {
-				input: 0,
-				output: 0,
-				cacheRead: 0,
-				cacheWrite: 0,
-			},
-			contextWindow: 262144,
-			maxTokens: 32000,
-		} satisfies Model<"openai-completions">,
-		"kimi-k2.5": {
-			id: "kimi-k2.5",
-			name: "Kimi K2.5",
-			api: "openai-completions",
-			provider: "kimi-code",
-			baseUrl: "https://api.kimi.com/coding/v1",
-			headers: {"User-Agent":"KimiCLI/1.0","X-Msh-Platform":"kimi_cli"},
-			compat: {"thinkingFormat":"zai","reasoningContentField":"reasoning_content","supportsDeveloperRole":false},
-			reasoning: true,
-			input: ["text", "image"],
-			cost: {
-				input: 0,
-				output: 0,
-				cacheRead: 0,
-				cacheWrite: 0,
-			},
-			contextWindow: 262144,
-			maxTokens: 32000,
-		} satisfies Model<"openai-completions">,
 	},
 	"minimax": {
 		"MiniMax-M2": {
@@ -4030,6 +4094,23 @@ export const MODELS = {
 			contextWindow: 400000,
 			maxTokens: 128000,
 		} satisfies Model<"openai-responses">,
+		"gpt-5.3-codex": {
+			id: "gpt-5.3-codex",
+			name: "GPT-5.3 Codex",
+			api: "openai-responses",
+			provider: "openai",
+			baseUrl: "https://api.openai.com/v1",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 1.75,
+				output: 14,
+				cacheRead: 0.175,
+				cacheWrite: 0,
+			},
+			contextWindow: 400000,
+			maxTokens: 128000,
+		} satisfies Model<"openai-responses">,
 		"o1": {
 			id: "o1",
 			name: "o1",
@@ -4253,6 +4334,23 @@ export const MODELS = {
 			contextWindow: 272000,
 			maxTokens: 128000,
 		} satisfies Model<"openai-codex-responses">,
+		"gpt-5.3-codex": {
+			id: "gpt-5.3-codex",
+			name: "GPT-5.3 Codex",
+			api: "openai-codex-responses",
+			provider: "openai-codex",
+			baseUrl: "https://chatgpt.com/backend-api",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 1.75,
+				output: 14,
+				cacheRead: 0.175,
+				cacheWrite: 0,
+			},
+			contextWindow: 400000,
+			maxTokens: 128000,
+		} satisfies Model<"openai-codex-responses">,
 	},
 	"opencode": {
 		"big-pickle": {
@@ -4340,6 +4438,23 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 64000,
 		} satisfies Model<"anthropic-messages">,
+		"claude-opus-4-6": {
+			id: "claude-opus-4-6",
+			name: "Claude Opus 4.6",
+			api: "anthropic-messages",
+			provider: "opencode",
+			baseUrl: "https://opencode.ai/zen",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 0.5,
+				cacheWrite: 6.25,
+			},
+			contextWindow: 1000000,
+			maxTokens: 128000,
+		} satisfies Model<"anthropic-messages">,
 		"claude-sonnet-4": {
 			id: "claude-sonnet-4",
 			name: "Claude Sonnet 4",
@@ -5060,6 +5175,23 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 64000,
 		} satisfies Model<"openai-completions">,
+		"anthropic/claude-opus-4.6": {
+			id: "anthropic/claude-opus-4.6",
+			name: "Anthropic: Claude Opus 4.6",
+			api: "openai-completions",
+			provider: "openrouter",
+			baseUrl: "https://openrouter.ai/api/v1",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 0.5,
+				cacheWrite: 6.25,
+			},
+			contextWindow: 1000000,
+			maxTokens: 128000,
+		} satisfies Model<"openai-completions">,
 		"anthropic/claude-sonnet-4": {
 			id: "anthropic/claude-sonnet-4",
 			name: "Anthropic: Claude Sonnet 4",
@@ -5265,57 +5397,6 @@ export const MODELS = {
 			contextWindow: 128000,
 			maxTokens: 4000,
 		} satisfies Model<"openai-completions">,
-		"deepcogito/cogito-v2-preview-llama-109b-moe": {
-			id: "deepcogito/cogito-v2-preview-llama-109b-moe",
-			name: "Cogito V2 Preview Llama 109B",
-			api: "openai-completions",
-			provider: "openrouter",
-			baseUrl: "https://openrouter.ai/api/v1",
-			reasoning: true,
-			input: ["text", "image"],
-			cost: {
-				input: 0.18,
-				output: 0.59,
-				cacheRead: 0,
-				cacheWrite: 0,
-			},
-			contextWindow: 32767,
-			maxTokens: 4096,
-		} satisfies Model<"openai-completions">,
-		"deepcogito/cogito-v2-preview-llama-405b": {
-			id: "deepcogito/cogito-v2-preview-llama-405b",
-			name: "Deep Cogito: Cogito V2 Preview Llama 405B",
-			api: "openai-completions",
-			provider: "openrouter",
-			baseUrl: "https://openrouter.ai/api/v1",
-			reasoning: true,
-			input: ["text"],
-			cost: {
-				input: 3.5,
-				output: 3.5,
-				cacheRead: 0,
-				cacheWrite: 0,
-			},
-			contextWindow: 32768,
-			maxTokens: 4096,
-		} satisfies Model<"openai-completions">,
-		"deepcogito/cogito-v2-preview-llama-70b": {
-			id: "deepcogito/cogito-v2-preview-llama-70b",
-			name: "Deep Cogito: Cogito V2 Preview Llama 70B",
-			api: "openai-completions",
-			provider: "openrouter",
-			baseUrl: "https://openrouter.ai/api/v1",
-			reasoning: true,
-			input: ["text"],
-			cost: {
-				input: 0.88,
-				output: 0.88,
-				cacheRead: 0,
-				cacheWrite: 0,
-			},
-			contextWindow: 32768,
-			maxTokens: 4096,
-		} satisfies Model<"openai-completions">,
 		"deepseek/deepseek-chat": {
 			id: "deepseek/deepseek-chat",
 			name: "DeepSeek: DeepSeek V3",
@@ -5412,7 +5493,7 @@ export const MODELS = {
 			cost: {
 				input: 0.21,
 				output: 0.7899999999999999,
-				cacheRead: 0.16799999999999998,
+				cacheRead: 0.1300000002,
 				cacheWrite: 0,
 			},
 			contextWindow: 163840,
@@ -9042,6 +9123,23 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 64000,
 		} satisfies Model<"anthropic-messages">,
+		"anthropic/claude-opus-4.6": {
+			id: "anthropic/claude-opus-4.6",
+			name: "Claude Opus 4.6",
+			api: "anthropic-messages",
+			provider: "vercel-ai-gateway",
+			baseUrl: "https://ai-gateway.vercel.sh",
+			reasoning: true,
+			input: ["text", "image"],
+			cost: {
+				input: 5,
+				output: 25,
+				cacheRead: 0.5,
+				cacheWrite: 6.25,
+			},
+			contextWindow: 1000000,
+			maxTokens: 128000,
+		} satisfies Model<"anthropic-messages">,
 		"anthropic/claude-sonnet-4": {
 			id: "anthropic/claude-sonnet-4",
 			name: "Claude Sonnet 4",
@@ -9799,13 +9897,13 @@ export const MODELS = {
 			reasoning: true,
 			input: ["text", "image"],
 			cost: {
-				input: 0.44999999999999996,
+				input: 0.5,
 				output: 2.8,
 				cacheRead: 0,
 				cacheWrite: 0,
 			},
-			contextWindow: 262144,
-			maxTokens: 252144,
+			contextWindow: 256000,
+			maxTokens: 256000,
 		} satisfies Model<"anthropic-messages">,
 		"nvidia/nemotron-nano-12b-v2-vl": {
 			id: "nvidia/nemotron-nano-12b-v2-vl",

package/src/providers/google-gemini-cli.ts

@@ -5,7 +5,7 @@
  */
 import { createHash } from "node:crypto";
 import type { Content, ThinkingConfig } from "@google/genai";
-import { abortableSleep } from "@oh-my-pi/pi-utils";
+import { abortableSleep, readSseJson } from "@oh-my-pi/pi-utils";
 import { calculateCost } from "../models";
 import type {
 	Api,
@@ -523,211 +523,168 @@ export const streamGoogleGeminiCli: StreamFunction<"google-gemini-cli"> = (
 				const blocks = output.content;
 				const blockIndex = () => blocks.length - 1;
 
-				// Read SSE stream
-				const reader = activeResponse.body.getReader();
-				const decoder = new TextDecoder();
-				let buffer = "";
-				let jsonlBuffer = "";
-
-				// Set up abort handler to cancel reader when signal fires
-				const abortHandler = () => {
-					void reader.cancel().catch(() => {});
-				};
-				options?.signal?.addEventListener("abort", abortHandler);
-
-				try {
-					while (true) {
-						// Check abort signal before each read
-						if (options?.signal?.aborted) {
-							throw new Error("Request was aborted");
-						}
-
-						const { done, value } = await reader.read();
-						if (done) break;
-
-						buffer += decoder.decode(value, { stream: true });
-						const lines = buffer.split("\n");
-						buffer = lines.pop() || "";
-
-						for (const line of lines) {
-							if (!line.startsWith("data:")) continue;
-
-							const jsonStr = line.slice(5).trim();
-							if (!jsonStr) continue;
-							jsonlBuffer += `${jsonStr}\n`;
-							const parsed = Bun.JSONL.parseChunk(jsonlBuffer);
-							jsonlBuffer = jsonlBuffer.slice(parsed.read);
-							if (parsed.error) {
-								jsonlBuffer = "";
-								continue;
-							}
-
-							const chunk = parsed.values[0] as CloudCodeAssistResponseChunk | undefined;
-							if (!chunk) continue;
-
-							// Unwrap the response
-							const responseData = chunk.response;
-							if (!responseData) continue;
-
-							const candidate = responseData.candidates?.[0];
-							if (candidate?.content?.parts) {
-								for (const part of candidate.content.parts) {
-									if (part.text !== undefined) {
-										hasContent = true;
-										const isThinking = isThinkingPart(part);
-										if (
-											!currentBlock ||
-											(isThinking && currentBlock.type !== "thinking") ||
-											(!isThinking && currentBlock.type !== "text")
-										) {
-											if (currentBlock) {
-												if (currentBlock.type === "text") {
-													stream.push({
-														type: "text_end",
-														contentIndex: blocks.length - 1,
-														content: currentBlock.text,
-														partial: output,
-													});
-												} else {
-													stream.push({
-														type: "thinking_end",
-														contentIndex: blockIndex(),
-														content: currentBlock.thinking,
-														partial: output,
-													});
-												}
-											}
-											if (isThinking) {
-												currentBlock = { type: "thinking", thinking: "", thinkingSignature: undefined };
-												output.content.push(currentBlock);
-												ensureStarted();
-												stream.push({
-													type: "thinking_start",
-													contentIndex: blockIndex(),
-													partial: output,
-												});
-											} else {
-												currentBlock = { type: "text", text: "" };
-												output.content.push(currentBlock);
-												ensureStarted();
-												stream.push({ type: "text_start", contentIndex: blockIndex(), partial: output });
-											}
-										}
-										if (currentBlock.type === "thinking") {
-											currentBlock.thinking += part.text;
-											currentBlock.thinkingSignature = retainThoughtSignature(
-												currentBlock.thinkingSignature,
-												part.thoughtSignature,
-											);
+				for await (const chunk of readSseJson<CloudCodeAssistResponseChunk>(
+					activeResponse.body!,
+					options?.signal,
+				)) {
+					const responseData = chunk.response;
+					if (!responseData) continue;
+
+					const candidate = responseData.candidates?.[0];
+					if (candidate?.content?.parts) {
+						for (const part of candidate.content.parts) {
+							if (part.text !== undefined) {
+								hasContent = true;
+								const isThinking = isThinkingPart(part);
+								if (
+									!currentBlock ||
+									(isThinking && currentBlock.type !== "thinking") ||
+									(!isThinking && currentBlock.type !== "text")
+								) {
+									if (currentBlock) {
+										if (currentBlock.type === "text") {
 											stream.push({
-												type: "thinking_delta",
-												contentIndex: blockIndex(),
-												delta: part.text,
+												type: "text_end",
+												contentIndex: blocks.length - 1,
+												content: currentBlock.text,
 												partial: output,
 											});
 										} else {
-											currentBlock.text += part.text;
-											currentBlock.textSignature = retainThoughtSignature(
-												currentBlock.textSignature,
-												part.thoughtSignature,
-											);
 											stream.push({
-												type: "text_delta",
+												type: "thinking_end",
 												contentIndex: blockIndex(),
-												delta: part.text,
+												content: currentBlock.thinking,
 												partial: output,
 											});
 										}
 									}
-
-									if (part.functionCall) {
-										hasContent = true;
-										if (currentBlock) {
-											if (currentBlock.type === "text") {
-												stream.push({
-													type: "text_end",
-													contentIndex: blockIndex(),
-													content: currentBlock.text,
-													partial: output,
-												});
-											} else {
-												stream.push({
-													type: "thinking_end",
-													contentIndex: blockIndex(),
-													content: currentBlock.thinking,
-													partial: output,
-												});
-											}
-											currentBlock = null;
-										}
-
-										const providedId = part.functionCall.id;
-										const needsNewId =
-											!providedId || output.content.some(b => b.type === "toolCall" && b.id === providedId);
-										const toolCallId = needsNewId
-											? `${part.functionCall.name}_${Date.now()}_${++toolCallCounter}`
-											: providedId;
-
-										const toolCall: ToolCall = {
-											type: "toolCall",
-											id: toolCallId,
-											name: part.functionCall.name || "",
-											arguments: part.functionCall.args as Record<string, unknown>,
-											...(part.thoughtSignature && { thoughtSignature: part.thoughtSignature }),
-										};
-
-										output.content.push(toolCall);
+									if (isThinking) {
+										currentBlock = { type: "thinking", thinking: "", thinkingSignature: undefined };
+										output.content.push(currentBlock);
 										ensureStarted();
-										stream.push({ type: "toolcall_start", contentIndex: blockIndex(), partial: output });
 										stream.push({
-											type: "toolcall_delta",
+											type: "thinking_start",
 											contentIndex: blockIndex(),
-											delta: JSON.stringify(toolCall.arguments),
 											partial: output,
 										});
+									} else {
+										currentBlock = { type: "text", text: "" };
+										output.content.push(currentBlock);
+										ensureStarted();
+										stream.push({ type: "text_start", contentIndex: blockIndex(), partial: output });
+									}
+								}
+								if (currentBlock.type === "thinking") {
+									currentBlock.thinking += part.text;
+									currentBlock.thinkingSignature = retainThoughtSignature(
+										currentBlock.thinkingSignature,
+										part.thoughtSignature,
+									);
+									stream.push({
+										type: "thinking_delta",
+										contentIndex: blockIndex(),
+										delta: part.text,
+										partial: output,
+									});
+								} else {
+									currentBlock.text += part.text;
+									currentBlock.textSignature = retainThoughtSignature(
+										currentBlock.textSignature,
+										part.thoughtSignature,
+									);
+									stream.push({
+										type: "text_delta",
+										contentIndex: blockIndex(),
+										delta: part.text,
+										partial: output,
+									});
+								}
+							}
+
+							if (part.functionCall) {
+								hasContent = true;
+								if (currentBlock) {
+									if (currentBlock.type === "text") {
 										stream.push({
-											type: "toolcall_end",
+											type: "text_end",
 											contentIndex: blockIndex(),
-											toolCall,
+											content: currentBlock.text,
+											partial: output,
+										});
+									} else {
+										stream.push({
+											type: "thinking_end",
+											contentIndex: blockIndex(),
+											content: currentBlock.thinking,
 											partial: output,
 										});
 									}
+									currentBlock = null;
 								}
-							}
-
-							if (candidate?.finishReason) {
-								output.stopReason = mapStopReasonString(candidate.finishReason);
-								if (output.content.some(b => b.type === "toolCall")) {
-									output.stopReason = "toolUse";
-								}
-							}
 
-							if (responseData.usageMetadata) {
-								// promptTokenCount includes cachedContentTokenCount, so subtract to get fresh input
-								const promptTokens = responseData.usageMetadata.promptTokenCount || 0;
-								const cacheReadTokens = responseData.usageMetadata.cachedContentTokenCount || 0;
-								output.usage = {
-									input: promptTokens - cacheReadTokens,
-									output:
-										(responseData.usageMetadata.candidatesTokenCount || 0) +
-										(responseData.usageMetadata.thoughtsTokenCount || 0),
-									cacheRead: cacheReadTokens,
-									cacheWrite: 0,
-									totalTokens: responseData.usageMetadata.totalTokenCount || 0,
-									cost: {
-										input: 0,
-										output: 0,
-										cacheRead: 0,
-										cacheWrite: 0,
-										total: 0,
-									},
+								const providedId = part.functionCall.id;
+								const needsNewId =
+									!providedId || output.content.some(b => b.type === "toolCall" && b.id === providedId);
+								const toolCallId = needsNewId
+									? `${part.functionCall.name}_${Date.now()}_${++toolCallCounter}`
+									: providedId;
+
+								const toolCall: ToolCall = {
+									type: "toolCall",
+									id: toolCallId,
+									name: part.functionCall.name || "",
+									arguments: part.functionCall.args as Record<string, unknown>,
+									...(part.thoughtSignature && { thoughtSignature: part.thoughtSignature }),
 								};
-								calculateCost(model, output.usage);
+
+								output.content.push(toolCall);
+								ensureStarted();
+								stream.push({ type: "toolcall_start", contentIndex: blockIndex(), partial: output });
+								stream.push({
+									type: "toolcall_delta",
+									contentIndex: blockIndex(),
+									delta: JSON.stringify(toolCall.arguments),
+									partial: output,
+								});
+								stream.push({
+									type: "toolcall_end",
+									contentIndex: blockIndex(),
+									toolCall,
+									partial: output,
+								});
 							}
 						}
 					}
-				} finally {
-					options?.signal?.removeEventListener("abort", abortHandler);
+
+					if (candidate?.finishReason) {
+						output.stopReason = mapStopReasonString(candidate.finishReason);
+						if (output.content.some(b => b.type === "toolCall")) {
+							output.stopReason = "toolUse";
+						}
+					}
+
+					if (responseData.usageMetadata) {
+						// promptTokenCount includes cachedContentTokenCount, so subtract to get fresh input
+						const promptTokens = responseData.usageMetadata.promptTokenCount || 0;
+						const cacheReadTokens = responseData.usageMetadata.cachedContentTokenCount || 0;
+						output.usage = {
+							input: promptTokens - cacheReadTokens,
+							output:
+								(responseData.usageMetadata.candidatesTokenCount || 0) +
+								(responseData.usageMetadata.thoughtsTokenCount || 0),
+							cacheRead: cacheReadTokens,
+							cacheWrite: 0,
+							totalTokens: responseData.usageMetadata.totalTokenCount || 0,
+							cost: {
+								input: 0,
+								output: 0,
+								cacheRead: 0,
+								cacheWrite: 0,
+								total: 0,
+							},
+						};
+						calculateCost(model, output.usage);
+					}
 				}
 
 				if (currentBlock) {