<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
<channel>
  <title>AI Radar · 开源</title>
  <link>https://ai-radar-ik42.pages.dev/</link>
  <description>AI Radar 自动聚合 AI 官方博客、GitHub Release、可信媒体与社区讨论，将同一事件的多个来源聚类，标注验证状态（已确认 / 多源佐证 / 有限来源 / 传闻），并持续追踪后续进展。</description>
  <language>zh-cn</language>
  <item>
    <title>[已确认] llama.cpp b10481: CUDA: MMVQ nwarps=8 for bs=1 for dense models on DGX Spark (#26843)</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-f5e725a80e9e/index.html</link>
    <guid isPermaLink="false">ev-f5e725a80e9e-zh</guid>
    <pubDate>Tue, 18 Aug 2026 04:15:53 GMT</pubDate>
    <description>已确认 · 1 src · • CUDA: MMVQ nwarps=8 for bs=1 for dense models on DGX Spark

Signed-off-by: ynankani ynankani@nvidia.com

• skip moe experts and allow others based on k geometry (allow only small idle tail)

Signed-off-by: ynankani ynankani@nvidia.com

• rename MMVQ DGX Spark params to GB10 an…</description>
  </item>
  <item>
    <title>[有限来源] 本地 Agentic 编程评测：Qwen 3.8 27B（多种权重量化/缓存量化/引擎/推理档位）对比其他模型</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-f32b2a2af121/index.html</link>
    <guid isPermaLink="false">ev-f32b2a2af121-zh</guid>
    <pubDate>Mon, 17 Aug 2026 20:45:23 GMT</pubDate>
    <description>有限来源 · 1 src · In medium reasoning mode, it both scores higher than the 3.6 version, AND is very much more efficient (almost half requests needed, and a third less tokens generated) - at DeepSeek v4 Flash 3107 MXFP4 level
The xhigh mode is advertised to be the best one for hard tasks. In this…</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.234</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-b4118147a267/index.html</link>
    <guid isPermaLink="false">ev-b4118147a267-zh</guid>
    <pubDate>Mon, 17 Aug 2026 20:20:58 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Added the optional CLAUDE_CODE_PROJECT_DIR_NAME environment variable: hosts that give each session its own config directory can choose a short name for the per-project transcript directory
Added the selection:clear keybinding action, so a key can be bound to clear…</description>
  </item>
  <item>
    <title>[已确认] openai-python v3.2.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-10d44cb0b131/index.html</link>
    <guid isPermaLink="false">ev-10d44cb0b131-zh</guid>
    <pubDate>Mon, 17 Aug 2026 19:13:27 GMT</pubDate>
    <description>已确认 · 1 src · 3.2.0 (2026-08-17)
Features
add Bedrock Runtime endpoint support (SDK-290) ( #3623 ) ( 86267d2 )
api: Add shell call streaming events and new service/image types ( #3635 ) ( ff14a33 )</description>
  </item>
  <item>
    <title>[已确认] llama.cpp 发布 v0.1.0：本地推理项目进入正式版本里程碑</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-4033aa830e01/index.html</link>
    <guid isPermaLink="false">ev-4033aa830e01-zh</guid>
    <pubDate>Mon, 17 Aug 2026 09:08:00 GMT</pubDate>
    <description>已确认 · 3 src · Release v0.1.0</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10472</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-afc566416f09/index.html</link>
    <guid isPermaLink="false">ev-afc566416f09-zh</guid>
    <pubDate>Mon, 17 Aug 2026 17:18:33 GMT</pubDate>
    <description>已确认 · 1 src · cuda : skip UMA override for HIP builds ( #27083 )
AMD APUs report accurate memory via hipMemGetInfo. Using
MemAvailable over-promises on small-carveout systems.
fixes #18159
Website:
https://llama.app
macOS/iOS:
macOS Apple Silicon (arm64)
macOS Apple Silicon (arm64, KleidiAI e…</description>
  </item>
  <item>
    <title>[已确认] llama.cpp v0.1.1</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-bf877550967d/index.html</link>
    <guid isPermaLink="false">ev-bf877550967d-zh</guid>
    <pubDate>Mon, 17 Aug 2026 14:08:29 GMT</pubDate>
    <description>已确认 · 1 src · Release v0.1.1</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10470</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-cc0230071414/index.html</link>
    <guid isPermaLink="false">ev-cc0230071414-zh</guid>
    <pubDate>Mon, 17 Aug 2026 14:01:13 GMT</pubDate>
    <description>已确认 · 1 src · ci : push release tag explicitly in release.yml ( #27261 )
Add a &quot;Create and push git tag&quot; step to the release job, right before
the &quot;Create release&quot; step. The tag is created with git tag and pushed
with the deploy key already configured by the Clone step, instead of
relying on…</description>
  </item>
  <item>
    <title>[有限来源] GPT 5.6 Sol 是 OpenAI 发布过的最强「视觉」模型</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-b13c8a2e3d0c/index.html</link>
    <guid isPermaLink="false">ev-b13c8a2e3d0c-zh</guid>
    <pubDate>Mon, 17 Aug 2026 12:09:42 GMT</pubDate>
    <description>有限来源 · 1 src · Article URL: https://blog.roboflow.com/openai-gpt-5-6/
Comments URL: https://news.ycombinator.com/item?id=49329575
Points: 149
# Comments: 82</description>
  </item>
  <item>
    <title>[有限来源] [主题楼] Qwen 3.8 27B 发布日</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-d92bad1edd18/index.html</link>
    <guid isPermaLink="false">ev-d92bad1edd18-zh</guid>
    <pubDate>Sat, 15 Aug 2026 00:41:13 GMT</pubDate>
    <description>有限来源 · 2 src · Megathread to help with the influx of duplicate / similar posts around the release of the Qwen 3.8 27B release.
Quants
Fine-Tunes &amp; Abliterations
Chat Templates
Inference Server Support &amp; Configuration
Experiences, Benchmarks &amp; Model Comparisons
Official:
https://huggingface.co/…</description>
  </item>
  <item>
    <title>[已确认] llama.cpp tmp-testing-0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-f68304b9f67f/index.html</link>
    <guid isPermaLink="false">ev-f68304b9f67f-zh</guid>
    <pubDate>Mon, 17 Aug 2026 06:58:11 GMT</pubDate>
    <description>已确认 · 1 src · ci : make release workflows use a deply key</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10456</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-b401282dcb6a/index.html</link>
    <guid isPermaLink="false">ev-b401282dcb6a-zh</guid>
    <pubDate>Mon, 17 Aug 2026 06:31:36 GMT</pubDate>
    <description>已确认 · 1 src · sycl: fix thread/block count in quantized cpy kernel launches ( #27160 )
Adjusts the thread/block count to be proportional to the size
of the quant, reducing under/over subscription.
Largest perf improvement is the q4_0 -&gt; f32 path, with, on
a Arc 70, throughput goes from 20.21…</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10455</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-bccfa2e2bd60/index.html</link>
    <guid isPermaLink="false">ev-bccfa2e2bd60-zh</guid>
    <pubDate>Mon, 17 Aug 2026 05:49:33 GMT</pubDate>
    <description>已确认 · 1 src · [SYCL] support OP OPT_STEP_ADAMW, OPT_STEP_SGD ( #25268 )
fix conflict
fix conflict of ops.md
fix conflict of ops.md
update the ops.md
Co-authored-by: Neo Zhang Jianyu jianyu.zhang@intel.com
Website:
https://llama.app
macOS/iOS:
macOS Apple Silicon (arm64)
macOS Apple Silicon (a…</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.14</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-b7364f128fc7/index.html</link>
    <guid isPermaLink="false">ev-b7364f128fc7-zh</guid>
    <pubDate>Sun, 16 Aug 2026 17:41:11 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s Changed
llm: transcode WebP images for llama-server
renderers/qwen: tolerate non-leading system messages
Full Changelog : v0.32.13...v0.32.14</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10453</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-3fe1e1b19b81/index.html</link>
    <guid isPermaLink="false">ev-3fe1e1b19b81-zh</guid>
    <pubDate>Sun, 16 Aug 2026 12:55:52 GMT</pubDate>
    <description>已确认 · 1 src · model : remove some ggml_concat ( #27176 )
Co-authored-by: Xuan Son Nguyen son@huggingface.co
Website:
https://llama.app
macOS/iOS:
macOS Apple Silicon (arm64)
macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED
macOS Intel (x64)
iOS XCFramework
Linux:
Ubuntu x64 (CPU)
Ubuntu…</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10454: ci : fix dry-run reporting in make-release job [no ci] (#27167)</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-fd26c85fe2cc/index.html</link>
    <guid isPermaLink="false">ev-fd26c85fe2cc-zh</guid>
    <pubDate>Sun, 16 Aug 2026 12:53:13 GMT</pubDate>
    <description>已确认 · 1 src · This commit fixes the reporting in the make-release CI job when
--dry-run is used. It will currently incorrectly report that all checks
pass even if there are steps that fail.
Refs: #26839 (comment)</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10452</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-74755e034613/index.html</link>
    <guid isPermaLink="false">ev-74755e034613-zh</guid>
    <pubDate>Sun, 16 Aug 2026 11:23:15 GMT</pubDate>
    <description>已确认 · 1 src · chat: refactor handling supports_string_content / supports_typed_content ( #27130 )
better supports_string_content cap detect
test: add &quot;skip&quot;
messages_inp_normalizer
Website:
https://llama.app
macOS/iOS:
macOS Apple Silicon (arm64)
macOS Apple Silicon (arm64, KleidiAI enabled)…</description>
  </item>
  <item>
    <title>[已确认] llama.cpp b10451</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-b94c31554760/index.html</link>
    <guid isPermaLink="false">ev-b94c31554760-zh</guid>
    <pubDate>Sun, 16 Aug 2026 07:26:01 GMT</pubDate>
    <description>已确认 · 1 src · llama : check LoRA tensor data is within file bounds ( #27056 )
llama : check LoRA tensor data is within file bounds
Update src/llama-adapter.cpp
Co-authored-by: Sigbjørn Skjæret sigbjorn.skjaeret@huggingface.co
Co-authored-by: Sigbjørn Skjæret sigbjorn.skjaeret@huggingface.co
W…</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.14-rc0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-147f2f42c214/index.html</link>
    <guid isPermaLink="false">ev-147f2f42c214-zh</guid>
    <pubDate>Sat, 15 Aug 2026 18:56:40 GMT</pubDate>
    <description>已确认 · 1 src · mlx update ( #17761 )</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.12</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-486e4bd3a5ae/index.html</link>
    <guid isPermaLink="false">ev-486e4bd3a5ae-zh</guid>
    <pubDate>Sat, 15 Aug 2026 04:30:53 GMT</pubDate>
    <description>已确认 · 1 src · Qwen 3.8 27B
This release adds the support of Qwen 3.8 27B . Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks.
ollama run qwen3.8:27b
For Apple Silicon devices, Ollama has in particular optimized for maximum performanc…</description>
  </item>
  <item>
    <title>[已确认] openai-python v3.1.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-acfd616f4cdf/index.html</link>
    <guid isPermaLink="false">ev-acfd616f4cdf-zh</guid>
    <pubDate>Fri, 14 Aug 2026 23:48:55 GMT</pubDate>
    <description>已确认 · 1 src · 3.1.0 (2026-08-14)
Features
api: add WebSocket stream IDs ( #3612 ) ( d9029e3 )
api: add workload identity access token issued event ( #3601 ) ( df274d4 )
api: deprecate Sora video APIs ( #3610 ) ( 721cb1c )
api: Ultrafast tier, structured MCP and websocket errors, separate webs…</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.233</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-78e1f6c011cf/index.html</link>
    <guid isPermaLink="false">ev-78e1f6c011cf-zh</guid>
    <pubDate>Fri, 14 Aug 2026 22:20:57 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Added GitLab merge request URL support to the --worktree flag and the claude agents view (where MRs display as !N )
Added an opt-in forward_user_identity apps gateway setting on Anthropic upstreams that sends the signed-in user&#39;s identity as headers, so a proxy be…</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.13</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-30a44ff5b357/index.html</link>
    <guid isPermaLink="false">ev-30a44ff5b357-zh</guid>
    <pubDate>Fri, 14 Aug 2026 19:18:00 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s Changed
qwen3.8: support developer instructions
Full Changelog : v0.32.12...v0.32.13</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.11</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-675097889342/index.html</link>
    <guid isPermaLink="false">ev-675097889342-zh</guid>
    <pubDate>Fri, 14 Aug 2026 06:37:49 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s Changed
ollama launch dsh now supports DeepSeek Harness, DeepSeek&#39;s open-source agent harness
ollama launch muse now supports Muse Code , Meta&#39;s agentic coding CLI
The OpenAI-compatible Responses API now supports web search
Muse Glimmer template updates
Full Changelog : v…</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.232</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-1fd438e0fc8b/index.html</link>
    <guid isPermaLink="false">ev-1fd438e0fc8b-zh</guid>
    <pubDate>Thu, 13 Aug 2026 23:29:59 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Subagent forking is now on by default: a subagent_type: &quot;fork&quot; subagent inherits the full conversation and prompt cache, and non-teammate agent spawns in interactive sessions now run in the background by default
Type @ in the prompt to mention another Claude sessi…</description>
  </item>
  <item>
    <title>[已确认] anthropic-sdk-python v0.122.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-75b6472bc0ba/index.html</link>
    <guid isPermaLink="false">ev-75b6472bc0ba-zh</guid>
    <pubDate>Thu, 13 Aug 2026 18:35:49 GMT</pubDate>
    <description>已确认 · 1 src · 0.122.0 (2026-08-13)
Full Changelog: v0.121.0...v0.122.0
Features
api: add output_behavior to dream creation (create a new memory store or update the input store in place) ( 852c4bb )
Bug Fixes
bedrock,aws: run SigV4 signing off the event loop in async clients ( #334 ) ( 2bae6c8…</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.10</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-f31b38a48c0d/index.html</link>
    <guid isPermaLink="false">ev-f31b38a48c0d-zh</guid>
    <pubDate>Thu, 13 Aug 2026 18:32:55 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s Changed
Models that don&#39;t set a repeat_penalty now default to 1.0 (off) instead of 1.1, matching other engines and speeding up speculative decoding; set a per-model parameter if an older model repeats itself.
Faster prefill on NVFP4 MLX models with a global scale, about 7…</description>
  </item>
  <item>
    <title>[已确认] Gemini 3.7 Flash 正式发布</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-5d2c0036a42b/index.html</link>
    <guid isPermaLink="false">ev-5d2c0036a42b-zh</guid>
    <pubDate>Thu, 13 Aug 2026 17:00:13 GMT</pubDate>
    <description>已确认 · 2 src · Google is announcing a new Gemini model today, but it&#39;s not the long-awaited 3.5 Pro. Gemini 3.7 Flash is now rolling out to replace 3.6 Flash, which itself was released only three weeks ago . This new &quot;workhorse&quot; model is supposedly the product of core optimizations and develop…</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.231</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-87be64d6bc53/index.html</link>
    <guid isPermaLink="false">ev-87be64d6bc53-zh</guid>
    <pubDate>Thu, 13 Aug 2026 08:38:50 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Fixed MCP OAuth sign-in failing with a redirect URI mismatch for servers that use a pre-registered OAuth client, such as Slack</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.10-rc1: mlx: avoid pulling MLX models when MLX is missing (#17710)</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-92c0752ba1af/index.html</link>
    <guid isPermaLink="false">ev-92c0752ba1af-zh</guid>
    <pubDate>Wed, 12 Aug 2026 21:42:17 GMT</pubDate>
    <description>已确认 · 1 src · As we look to bring Linux and Windows MLX support online, instead of blocking
downloads at the registry to avoid users wasting time downloading a model they
can&#39;t run, shift the logic to the local side which knows if MLX is present or not.</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.229</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-61d711302cdf/index.html</link>
    <guid isPermaLink="false">ev-61d711302cdf-zh</guid>
    <pubDate>Wed, 12 Aug 2026 20:56:22 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Documented claude remote-control --continue for resuming the most recent Remote Control session
Added server-supplied Claude Code hook support for self-hosted runner sessions, matching managed-environment behavior
Added SSE keepalive pings to gateway streaming res…</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.10-rc0: nn: speed up prefill on double-scale nvfp4 models</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-e364eb23c285/index.html</link>
    <guid isPermaLink="false">ev-e364eb23c285-zh</guid>
    <pubDate>Wed, 12 Aug 2026 20:25:33 GMT</pubDate>
    <description>已确认 · 1 src · ModelOpt checkpoints apply a float32 global scale to every projection
output on top of the per-group quantization scales. Running the
multiply and the cast back to the activation dtype as separate eager
ops costs an extra kernel launch and a materialized intermediate per
project…</description>
  </item>
  <item>
    <title>[已确认] vllm v0.27.2rc0: [Spec Decode] DSpark confidence-scheduled verification (#47808)</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-650ca3dbc713/index.html</link>
    <guid isPermaLink="false">ev-650ca3dbc713-zh</guid>
    <pubDate>Wed, 12 Aug 2026 19:27:48 GMT</pubDate>
    <description>已确认 · 1 src · Signed-off-by: Lucas Wilkinson lwilkins@redhat.com
Signed-off-by: Lucas Wilkinson LucasWilkinson@users.noreply.github.com
Signed-off-by: Benjamin Chislett chislett.ben@gmail.com
Signed-off-by: Lucas Wilkinson wilkinson.lucas@gmail.com
Signed-off-by: Nick Hill nickhill123@gmail.c…</description>
  </item>
  <item>
    <title>[已确认] openai-python v3.0.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-68de471dd313/index.html</link>
    <guid isPermaLink="false">ev-68de471dd313-zh</guid>
    <pubDate>Wed, 12 Aug 2026 01:54:46 GMT</pubDate>
    <description>已确认 · 1 src · 3.0.0 (2026-08-12)
⚠ BREAKING CHANGES
api: HTTPX2 is now the default HTTP client, and httpx is no longer installed automatically. Applications using custom HTTPX clients, transports, or configuration objects must migrate to their HTTPX2 equivalents or use the temporary, runtime-…</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.228</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-6ca674eab9d8/index.html</link>
    <guid isPermaLink="false">ev-6ca674eab9d8-zh</guid>
    <pubDate>Tue, 11 Aug 2026 19:50:59 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Fixed interactive sessions that could stop redrawing entirely, while the process kept running, after a rare internal layout error
Fixed git / Git Bash not being found on Windows when Claude Code is launched from a parent folder of the git installation
Fixed /tui r…</description>
  </item>
  <item>
    <title>[已确认] openai-python v2.54.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-7d499d6d00da/index.html</link>
    <guid isPermaLink="false">ev-7d499d6d00da-zh</guid>
    <pubDate>Tue, 11 Aug 2026 18:45:59 GMT</pubDate>
    <description>已确认 · 1 src · 2.54.0 (2026-08-11)
Features
api: Add new Responses model identifiers ( #3595 ) ( 0652787 )
Bug Fixes
api: clarify audio upload metadata requirements ( #3596 ) ( 28888f9 )
Chores
api: Update generated-file header attribution to Castiron ( #3583 ) ( ea17fda )</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.9</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-a521144b5e14/index.html</link>
    <guid isPermaLink="false">ev-a521144b5e14-zh</guid>
    <pubDate>Tue, 11 Aug 2026 13:24:14 GMT</pubDate>
    <description>已确认 · 1 src · NVIDIA Nemotron 3.5 Lightning
NVIDIA Nemotron 3.5 Lightning is an open 30B mixture-of-experts (MoE) model with 3B active parameters built for that execution layer of always-on agents. It is designed for harnesses like OpenClaw and Hermes Agent – all supported by the NVIDIA NemoC…</description>
  </item>
  <item>
    <title>[已确认] vllm v0.27.1</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-638fde7206c4/index.html</link>
    <guid isPermaLink="false">ev-638fde7206c4-zh</guid>
    <pubDate>Tue, 11 Aug 2026 10:49:10 GMT</pubDate>
    <description>已确认 · 1 src · This is a patch release on top of v0.27.0.
Support quantized DSpark Markov heads ( #50424 )</description>
  </item>
  <item>
    <title>[已确认] ollama v0.32.8</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-220d28266dc3/index.html</link>
    <guid isPermaLink="false">ev-220d28266dc3-zh</guid>
    <pubDate>Tue, 11 Aug 2026 06:39:00 GMT</pubDate>
    <description>已确认 · 1 src · Muse Glimmer
Muse Glimmer is now available on all platforms. Muse Glimmer can power coding agent applications such as Claude Code, Codex, Pi and more, as well as long-running personal assistants such as OpenClaw and Hermes.
Ollama&#39;s MLX engine provides state-of-the-art performan…</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.227</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-c3e5e6e8b2dc/index.html</link>
    <guid isPermaLink="false">ev-c3e5e6e8b2dc-zh</guid>
    <pubDate>Mon, 10 Aug 2026 22:56:53 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Fixed feature flags being evaluated without the user&#39;s subscription tier when a session started with an expired login token, which could wrongly prompt Max plan users to enable usage credits for Fable
Fixed every Bash command failing under claude-code-action with…</description>
  </item>
  <item>
    <title>[已确认] vllm v0.27.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-7a61fd842b6f/index.html</link>
    <guid isPermaLink="false">ev-7a61fd842b6f-zh</guid>
    <pubDate>Mon, 10 Aug 2026 21:18:11 GMT</pubDate>
    <description>已确认 · 1 src · vLLM v0.27.0 Release Notes
Highlights
This release features 561 commits from 242 contributors (64 new)!
Kimi K3 support with a full stack landing in one release: core model files and kernels ( #50089 , #50000 ), Python ( #50093 ) and Rust ( #50104 ) frontends, AttnRes kernels (…</description>
  </item>
  <item>
    <title>[已确认] 用 NVIDIA Magpie TTS 构建低延迟多语言语音 Agent：开放权重、部署全可控</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-78790573f9e8/index.html</link>
    <guid isPermaLink="false">ev-78790573f9e8-zh</guid>
    <pubDate>Mon, 10 Aug 2026 16:25:36 GMT</pubDate>
    <description>已确认 · 1 src · </description>
  </item>
  <item>
    <title>[已确认] transformers Release: v5.15.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-997ea9dcf004/index.html</link>
    <guid isPermaLink="false">ev-997ea9dcf004-zh</guid>
    <pubDate>Mon, 10 Aug 2026 10:28:13 GMT</pubDate>
    <description>已确认 · 1 src · Release v5.15.0
New Model additions
Meta Muse Glimmer
Muse Glimmer, released today, is Meta’s new multimodal model, especially designed for agentic use cases. Distilled from Muse to 30B parameters, and released under the Apache 2.0 license, it can be deployed to local setups for…</description>
  </item>
  <item>
    <title>[已确认] Meta 携 Muse Glimmer 回归：本地、Agentic、多模态、开源</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-e1583ed3dc68/index.html</link>
    <guid isPermaLink="false">ev-e1583ed3dc68-zh</guid>
    <pubDate>Mon, 10 Aug 2026 00:00:00 GMT</pubDate>
    <description>已确认 · 1 src · </description>
  </item>
  <item>
    <title>[已确认] vllm v0.27.0rc2</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-54d3331b95a6/index.html</link>
    <guid isPermaLink="false">ev-54d3331b95a6-zh</guid>
    <pubDate>Sun, 09 Aug 2026 09:57:05 GMT</pubDate>
    <description>已确认 · 1 src · v0.27.0rc2</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.226</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-2be25eb1e780/index.html</link>
    <guid isPermaLink="false">ev-2be25eb1e780-zh</guid>
    <pubDate>Sat, 08 Aug 2026 02:48:05 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Bug fixes and reliability improvements</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.225</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-3e2a99e1d698/index.html</link>
    <guid isPermaLink="false">ev-3e2a99e1d698-zh</guid>
    <pubDate>Sat, 08 Aug 2026 01:09:26 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Added gateway spend-limit support to Claude Code&#39;s usage warning; the limit-reached message now names the cap, its reset time, and the operator&#39;s message (requires the gateway on 2.1.225)
Added a workspace trust prompt to claude agents for untrusted directories, m…</description>
  </item>
  <item>
    <title>[已确认] anthropic-sdk-python v0.121.0</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-adb97e43d471/index.html</link>
    <guid isPermaLink="false">ev-adb97e43d471-zh</guid>
    <pubDate>Fri, 07 Aug 2026 17:10:51 GMT</pubDate>
    <description>已确认 · 1 src · 0.121.0 (2026-08-07)
Full Changelog: v0.120.2...v0.121.0
Features
api: add mid-conversation-tool-changes-2026-07-01 beta ( c7d1531 )
api: add support for session budgets, advisor tool, pinned inference location and skills auto-loading from GitHub ( 193bae0 )
Chores
api: remove r…</description>
  </item>
  <item>
    <title>[已确认] claude-code v2.1.224</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-66a2d753ac36/index.html</link>
    <guid isPermaLink="false">ev-66a2d753ac36-zh</guid>
    <pubDate>Fri, 07 Aug 2026 04:00:59 GMT</pubDate>
    <description>已确认 · 1 src · What&#39;s changed
Added self-hosted environments: claude self-hosted-runner turns your own machines or containers into a place Claude Code web, mobile, and desktop sessions can run, on Team and Enterprise plans
Added archive plugin source: install plugins from a zip over HTTPS with…</description>
  </item>
  <item>
    <title>[已确认] vllm v0.27.0rc1</title>
    <link>https://ai-radar-ik42.pages.dev/event/ev-84916ceb3b49/index.html</link>
    <guid isPermaLink="false">ev-84916ceb3b49-zh</guid>
    <pubDate>Fri, 07 Aug 2026 00:36:54 GMT</pubDate>
    <description>已确认 · 1 src · v0.27.0rc1</description>
  </item>
</channel>
</rss>
