diff --git a/packages/llm/llm-pi-ai/README.i18n.yaml b/packages/llm/llm-pi-ai/README.i18n.yaml
index 064203d767..a7e0c79077 100644
--- a/packages/llm/llm-pi-ai/README.i18n.yaml
+++ b/packages/llm/llm-pi-ai/README.i18n.yaml
@@ -87,8 +87,8 @@
   en: b6bd570c0159ba08
   zh: 7d2d91eea6a8d830
 /deepseek-ai-dsh-llm-pi-ai/known-limitations-and-deferred-work:
-  en: de6e60be33ddcd82
-  zh: dfd0f00289200b77
+  en: ee870ac2ac87decc
+  zh: 9e66c5e7400c4bc6
 /deepseek-ai-dsh-llm-pi-ai/known-limitations-and-deferred-work/dev-note:
   en: 3678dd0f6642c2a5
   zh: 75cbdd064cc64282
diff --git a/packages/llm/llm-pi-ai/README.md b/packages/llm/llm-pi-ai/README.md
index 3666cd251d..11a573ba1f 100644
--- a/packages/llm/llm-pi-ai/README.md
+++ b/packages/llm/llm-pi-ai/README.md
@@ -237,7 +237,7 @@ These limits define where the adapter stops and future work begins. They are cur
 - **Images are accepted only in user and tool-result messages** — images in system, developer, or assistant history fail with `UNSUPPORTED_CONTENT` on both conversion paths.
 - **Provider HTTP status is unavailable** — pi-ai error events do not expose a stable HTTP status across providers.
 - **Retry policy is provider-owned, not an SDK retry** — pi-ai SDK retries stay disabled so durable agent steps and `llm/retry` events own every visible attempt, and direct `ctx.llm.stream()` calls remain single-attempt.
-- **Streamed tool-call arguments are parsed once, when the call ends** — the installed pi-ai carries [`patches/@earendil-works__pi-ai@1.1.0.patch`](../../../patches/@earendil-works__pi-ai@1.1.0.patch), which removes the per-delta re-parse of the whole accumulated argument JSON in every stream adapter (upstream [earendil-works/pi#9265](https://github.com/earendil-works/pi/issues/9265)); unpatched, a multi-megabyte argument stream costs O(n²) CPU on the event loop and stalls every session in the process. Until `toolcall_end`, a pi-ai partial's tool-call `arguments` stays `{}`; this adapter reads only the delta strings and the finalized arguments. Re-apply or retire the patch on every pi-ai upgrade.
+- **Streamed JSON tool-call arguments avoid per-delta reparsing** — the installed pi-ai carries [`patches/@earendil-works__pi-ai@1.1.0.patch`](../../../patches/@earendil-works__pi-ai@1.1.0.patch), which removes the per-delta re-parse of the whole accumulated argument JSON in every stream adapter (upstream [earendil-works/pi#9265](https://github.com/earendil-works/pi/issues/9265)); unpatched, a multi-megabyte argument stream costs O(n²) CPU on the event loop and stalls every session in the process. Completion events supply parsed arguments; Responses may parse at both argument completion and output-item completion. This adapter reads only delta strings and finalized arguments. Re-apply or retire the patch on every pi-ai upgrade.
 
 <a id="dev-note"></a>
 ### Dev Note
diff --git a/packages/llm/llm-pi-ai/README.zh.md b/packages/llm/llm-pi-ai/README.zh.md
index 64cbf97155..b1dc29a4ab 100644
--- a/packages/llm/llm-pi-ai/README.zh.md
+++ b/packages/llm/llm-pi-ai/README.zh.md
@@ -237,7 +237,7 @@ pi-ai 事件变成 harness 的推理、文本、工具调用、用量与 finish
 - **图片只在 user 与工具结果消息中被接受**——system、developer 或 assistant 历史中的图片在两条转换路径上都会以 `UNSUPPORTED_CONTENT` 失败。
 - **提供方 HTTP 状态不可用**——pi-ai 错误事件不跨提供方暴露稳定 HTTP 状态。
 - **重试策略由提供方自有，而非 SDK 重试**——pi-ai SDK 重试保持禁用，因此持久 agent（智能体）步骤与 `llm/retry` 事件拥有每个可见尝试，直接 `ctx.llm.stream()` 调用仍是单次尝试。
-- **流式工具调用参数只在调用结束时解析一次**——安装的 pi-ai 带有 [`patches/@earendil-works__pi-ai@1.1.0.patch`](../../../patches/@earendil-works__pi-ai@1.1.0.patch)，它移除了每个流适配器中对整段累计参数 JSON 的逐 delta 重新解析（上游 [earendil-works/pi#9265](https://github.com/earendil-works/pi/issues/9265)）；未打补丁时，数 MB 的参数流会在事件循环上消耗 O(n²) CPU，并使进程内所有会话停滞。在 `toolcall_end` 之前，pi-ai partial 的工具调用 `arguments` 保持为 `{}`；本适配器只读取 delta 字符串与最终参数。每次升级 pi-ai 时都要重新应用或撤销该补丁。
+- **流式 JSON 工具调用参数避免逐 delta 重复解析**——安装的 pi-ai 带有 [`patches/@earendil-works__pi-ai@1.1.0.patch`](../../../patches/@earendil-works__pi-ai@1.1.0.patch)，它移除了每个流适配器中对整段累计参数 JSON 的逐 delta 重新解析（上游 [earendil-works/pi#9265](https://github.com/earendil-works/pi/issues/9265)）；未打补丁时，数 MB 的参数流会在事件循环上消耗 O(n²) CPU，并使进程内所有会话停滞。完成事件提供解析后的参数；Responses 可能在参数完成和输出项完成时分别解析。本适配器只读取 delta 字符串与最终参数。每次升级 pi-ai 时都要重新应用或撤销该补丁。
 
 <a id="dev-note"></a>
 ### 开发备注
diff --git a/packages/llm/llm-pi-ai/tests/catalog.spec.ts b/packages/llm/llm-pi-ai/tests/catalog.spec.ts
index 342c1d7703..edce032e47 100644
--- a/packages/llm/llm-pi-ai/tests/catalog.spec.ts
+++ b/packages/llm/llm-pi-ai/tests/catalog.spec.ts
@@ -442,13 +442,14 @@ describe('hand-declared providers', () => {
 describe('catalog routes with per-model configuration', () => {
   it('keeps both Azure protocols under the azure catalog route', () => {
     const provider = resolveProfiles({ azure: {} }).get('azure')?.piProvider
-    const models = provider?.getModels()
+    if (provider === undefined) throw new Error('the installed catalog ships no azure provider')
+    const models = provider.getModels()
+    const protocols = new Set(models.map(model => model.api))
 
-    expect(provider?.id).toBe('azure')
-    expect(models).toEqual(expect.arrayContaining([
-      expect.objectContaining({ provider: 'azure', api: 'azure-openai-responses', id: 'gpt-6.1-sol' }),
-      expect.objectContaining({ provider: 'azure', api: 'openai-completions', id: 'deepseek-v4-pro' }),
-    ]))
+    expect(provider.id).toBe('azure')
+    expect(protocols).toContain('azure-openai-responses')
+    expect(protocols).toContain('openai-completions')
+    expect(models.every(model => model.provider === 'azure')).toBe(true)
     expect(catalogProvider('azure-openai-responses')).toBeUndefined()
   })