chore(release): bump kimi-cli to 1.49.0 and kosong to 0.55.0 (#2503)

Co-authored-by: jackfish212 <jackfish212@outlook.com>

Author: sailist

CHANGELOG.md

@@ -11,7 +11,13 @@ Only write entries that are worth mentioning to users.
 
 ## Unreleased
 
+## 1.49.0 (2026-07-16)
+
+**Highlights**: The completion-token budget for Kimi providers now adapts to the model's remaining context window, reducing context-length overflow errors on long turns
+
+- LLM: Clamp the Kimi completion-token budget to the model's remaining context window — the CLI no longer sends a fixed `max_tokens=32000` but estimates the remaining context for each request and caps `max_completion_tokens` accordingly. Set the new `KIMI_MODEL_MAX_COMPLETION_TOKENS` env var for an explicit hard cap (`KIMI_MODEL_MAX_TOKENS` remains a compatibility alias; `0` or a negative value disables clamping)
 - Kosong: Stop Kimi from automatically sending the legacy `reasoning_effort` parameter when configuring thinking — requests now use `thinking.type` exclusively while preserving explicit legacy passthrough
+- Kosong: Fix empty-string `reasoning_content` from thinking models being dropped from history — a reply that reasoned but ended with empty reasoning was recorded as having no reasoning at all, so Preserved Thinking backends that require `reasoning_content` on every assistant message rejected the next request with a 400
 
 ## 1.47.0 (2026-06-05)
 

docs/en/release-notes/breaking-changes.md

@@ -4,6 +4,8 @@ This page documents breaking changes in Kimi Code CLI releases and provides migr
 
 ## Unreleased
 
+## 1.49.0
+
 ### Kimi no longer sends legacy `reasoning_effort` automatically
 
 Kimi thinking configuration now uses `thinking.type` exclusively. `Kimi.with_thinking(...)` no longer adds the legacy `reasoning_effort` parameter to requests.

docs/en/release-notes/changelog.md

@@ -4,7 +4,13 @@ This page documents the changes in each Kimi Code CLI release.
 
 ## Unreleased
 
+## 1.49.0 (2026-07-16)
+
+**Highlights**: The completion-token budget for Kimi providers now adapts to the model's remaining context window, reducing context-length overflow errors on long turns
+
+- LLM: Clamp the Kimi completion-token budget to the model's remaining context window — the CLI no longer sends a fixed `max_tokens=32000` but estimates the remaining context for each request and caps `max_completion_tokens` accordingly. Set the new `KIMI_MODEL_MAX_COMPLETION_TOKENS` env var for an explicit hard cap (`KIMI_MODEL_MAX_TOKENS` remains a compatibility alias; `0` or a negative value disables clamping)
 - Kosong: Stop Kimi from automatically sending the legacy `reasoning_effort` parameter when configuring thinking — requests now use `thinking.type` exclusively while preserving explicit legacy passthrough
+- Kosong: Fix empty-string `reasoning_content` from thinking models being dropped from history — a reply that reasoned but ended with empty reasoning was recorded as having no reasoning at all, so Preserved Thinking backends that require `reasoning_content` on every assistant message rejected the next request with a 400
 
 ## 1.47.0 (2026-06-05)
 

docs/zh/release-notes/breaking-changes.md

@@ -4,6 +4,8 @@
 
 ## 未发布
 
+## 1.49.0
+
 ### Kimi 不再自动发送旧版 `reasoning_effort`
 
 Kimi 的 Thinking 配置现在仅使用 `thinking.type`。`Kimi.with_thinking(...)` 不再自动把旧版 `reasoning_effort` 参数加入请求。

docs/zh/release-notes/changelog.md

@@ -4,7 +4,13 @@
 
 ## 未发布
 
+## 1.49.0 (2026-07-16)
+
+**亮点**：Kimi 供应商的补全 token 预算现在会根据模型剩余上下文窗口动态调整，减少长轮次中的上下文超限错误
+
+- LLM：将 Kimi 的补全 token 预算钳制在模型剩余上下文窗口内——CLI 不再固定发送 `max_tokens=32000`，而是按每次请求估算剩余上下文并据此设置 `max_completion_tokens` 上限。可通过新的环境变量 `KIMI_MODEL_MAX_COMPLETION_TOKENS` 设置显式硬上限（`KIMI_MODEL_MAX_TOKENS` 仍为兼容别名；设为 `0` 或负值可关闭钳制）
 - Kosong：配置 Thinking 模式时不再自动向 Kimi 请求发送旧版 `reasoning_effort` 参数——请求现在仅使用 `thinking.type`，同时保留显式传递旧版参数的兼容能力
+- Kosong：修复 Thinking 模型返回的空字符串 `reasoning_content` 被从历史中丢弃的问题——此前「思考过但内容为空」的回复会被当作「没有思考」，导致要求每条 Assistant 消息都携带 `reasoning_content` 的 Preserved Thinking 后端在下一次请求时返回 400
 
 ## 1.47.0 (2026-06-05)
 

packages/kimi-code/pyproject.toml

@@ -1,10 +1,10 @@
 [project]
 name = "kimi-code"
-version = "1.48.0"
+version = "1.49.0"
 description = "Kimi Code is a CLI agent that lives in your terminal."
 readme = "README.md"
 requires-python = ">=3.12"
-dependencies = ["kimi-cli==1.48.0"]
+dependencies = ["kimi-cli==1.49.0"]
 
 [project.scripts]
 kimi-code = "kimi_cli.__main__:main"

packages/kosong/CHANGELOG.md

@@ -2,8 +2,13 @@
 
 ## Unreleased
 
+## 0.55.0 (2026-07-16)
+
 - Kimi: Stop automatically sending the legacy `reasoning_effort` parameter when configuring thinking — requests now use `thinking.type` exclusively while preserving explicit legacy passthrough
 - Kimi: Preserve empty-string `reasoning_content` as `ThinkPart(think="")` in both streaming and non-streaming responses — previously the truthy check dropped empty deltas, conflating "reasoned but empty" with "no reasoning at all"; the stored (empty) ThinkPart is what makes `_convert_message` emit `reasoning_content` on the next request, so preserved-thinking backends that require the field on every assistant message no longer 400 after a reason-free turn
+- Kimi: Add `GenerationKwargs.max_completion_tokens` and normalize the deprecated `max_tokens` alias to it before requests — the implicit `max_tokens=32000` default is no longer sent when the field is unset, so the server (or a per-request override) decides the budget
+- Kimi: Accept per-request `generation_overrides` in `generate()` — callers such as kimi-cli can now clamp `max_completion_tokens` for a single request without rebuilding the provider; the Chaos provider forwards overrides to the wrapped provider when present
+- Core: Expose the `x-trace-id` response header as `StreamedMessage.trace_id` and fire the new `on_trace_id` callback on `generate()` as soon as the header is available; `APIStatusError` now also carries `trace_id`
 
 ## 0.53.0 (2026-04-28)
 

packages/kosong/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "kosong"
-version = "0.54.0"
+version = "0.55.0"
 description = "The LLM abstraction layer for modern AI agent applications."
 readme = "README.md"
 requires-python = ">=3.12"

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "kimi-cli"
-version = "1.48.0"
+version = "1.49.0"
 description = "Kimi Code CLI is your next CLI agent."
 readme = "README.md"
 requires-python = ">=3.12"
@@ -9,7 +9,7 @@ dependencies = [
     "aiofiles>=24.0,<26.0",
     "aiohttp==3.13.3",
     "typer==0.21.1",
-    "kosong[contrib]==0.54.0",
+    "kosong[contrib]==0.55.0",
     # loguru stays >=0.6.0 because notify-py (via batrachian-toad) caps it at <=0.6.0 on 3.14+.
     "loguru>=0.6.0,<0.8",
     "prompt-toolkit==3.0.52",