<h2>What's changed</h2> <ul> <li>Changed auto mode for Claude API and Enterprise users, and on Bedrock, Vertex, Foundry and gateways, to default to the server-side classifier, which does not charge for classifier overhead (<code>CLAUDE_CODE_AUTO_MODE_SERVER=0</code> opts out on B…
<h2>What's changed</h2> <ul> <li>Added AGENTS.md support: in a project with no CLAUDE.md, Claude Code reads AGENTS.md instead; change it under "Project instructions" in <code>/config</code> (not yet on Bedrock, Vertex or Foundry)</li> <li>Added <code>CLAUDE_GATEWAY_PROXY_IS_EGRES…
<h2>What's changed</h2> <ul> <li>Fixed every request failing with <code>400 … Input tag 'advisor_20260301'</code> when <code>ANTHROPIC_BASE_URL</code> points at a proxy or gateway (2.1.275 regression)</li> </ul>
<h2>What's Changed</h2> <ul> <li>Added first-run setup when running <code>ollama</code>, with options to sign in or continue locally. Setup completion is shared with the desktop app on macOS and Windows.</li> <li>Added <code>ollama://apps</code> to open the desktop app’s Apps pag…
<h2>What's changed</h2> <ul> <li>Added the signed-in account to Claude apps gateway sign-in: when the gateway names it, you confirm it before the credential is saved, and <code>/status</code> shows it</li> <li>Added a send-now key (ctrl+enter, or ctrl+x ctrl+s) that interrupts th…
<h2>What's Changed</h2> <ul> <li>Added first-run setup when running <code>ollama</code>, with options to sign in or continue locally. Setup completion is shared with the desktop app on macOS and Windows.</li> <li>Added <code>ollama://apps</code> to open the desktop app’s Apps pag…
<h2>What's changed</h2> <ul> <li>Added a visible warning when memory usage is critical, with steps to free memory or restart safely</li> <li>Added <code>CLAUDE_CODE_MCP_STARTUP_WAIT_MS</code> to bound how long the first non-interactive turn waits for connecting MCP servers (<code…
<h2>What's Changed</h2> <ul> <li>app: fix ChatGPT model selector spacing</li> <li>mlxrunner: Evict prefix cache snapshots from the active conversation</li> <li>mlxrunner: check system free memory and wait for evicted runners before loading the next MLX model</li> <li>llm: raise t…
<h2>What's Changed</h2> <ul> <li>app: fix ChatGPT model selector spacing</li> <li>mlxrunner: Evict prefix cache snapshots from the active conversation</li> <li>mlxrunner: check system free memory and wait for evicted runners before loading the next MLX model</li> <li>llm: raise t…
<h2>What's Changed</h2> <ul> <li>app: fix ChatGPT model selector spacing</li> <li>mlxrunner: Evict prefix cache snapshots from the active conversation</li> <li>mlxrunner: check system free memory and wait for evicted runners before loading the next MLX model</li> <li>llm: raise t…
<h2>What's changed</h2> <ul> <li>Added fast mode in Claude Code Remote sessions (cloud and self-hosted runners): the host's fast-mode setting or <code>/fast</code> typed in the session applies where your organization allows it</li> <li>Added mouse support to the <code>/config</co…
<h2>Overview</h2> <p>llama.cpp 0.4.1 adds Maple 20B-A1B, Tencent Hy 4, and Spark2.5 support, improves JSON schema handling, chat parsing, logging, and server child-process management, and updates ggml to v0.24.0.</p> <h3>API changes</h3> <ul> <li>Changed <code>llama_sampler_chain…
<h2>What's changed</h2> <ul> <li>Fixed read-only git commands in Bash unexpectedly asking for permission after a session had been running for a while (regression in 2.1.269)</li> </ul>
<h2>What's changed</h2> <ul> <li>Added <code>claude plugin eval</code>: run a plugin's eval suite against Claude Code and get scored, reproducible results (JSON + HTML report); see <code>claude plugin eval --help</code></li> <li>Added <code>/output-style [name]</code> to list and…
<h2>What's changed</h2> <ul> <li>Added to the Claude apps gateway: with <code>pricing:</code> set in <code>gateway.yaml</code>, signed-in Claude Code clients receive the same rates through managed settings, so <code>/cost</code> and telemetry match the spend meter</li> <li>Added …
<h2>Use Ollama models in ChatGPT Desktop</h2> <p>Ollama models can now be used directly in ChatGPT Desktop, so you can keep your existing workflow while running open models. Setup is available from the Ollama app on MacOS.</p> <p>This release also improves structured output perfo…
<h2>Use Ollama models in ChatGPT Desktop</h2> <p>Ollama models can now be used directly in ChatGPT Desktop, so you can keep your existing workflow while running open models. Setup is available from the Ollama app on MacOS.</p> <p>This release also improves structured output perfo…
<h2>What's changed</h2> <ul> <li>Added <code>maxEffortLevel</code> setting (top-level or per model under <code>modelSettings</code>): caps the effort level on every provider, including Bedrock, Vertex and Foundry; users can still pick a lower level</li> <li>Added <code>--system-p…
<h2>Use Ollama models in ChatGPT Desktop</h2> <p>Ollama models can now be used directly in ChatGPT Desktop, so you can keep your existing workflow while running open models. Setup is available from the Ollama app on MacOS.</p> <p>This release also improves structured output perfo…
<h2>Use Ollama models in ChatGPT Desktop</h2> <p>Ollama models can now be used directly in ChatGPT Desktop, so you can keep your existing workflow while running open models. Setup is available from the Ollama app on MacOS.</p> <p>This release also improves structured output perfo…
<h2>What's changed</h2> <ul> <li>Fixed a 2.1.265 regression affecting LLM-gateway and proxy setups: the undocumented <code>CLAUDE_CODE_USE_GATEWAY</code> environment variable, previously ignored unless <code>ANTHROPIC_BASE_URL</code> and <code>ANTHROPIC_AUTH_TOKEN</code> were bot…
<p>openai: finalize responses at the web search limit (<a class="issue-link js-issue-link" href="https://github.com/ollama/ollama/pull/18328">#18328</a>)</p>
<h2>What's changed</h2> <ul> <li>Added <code>user.email</code> and <code>user.groups</code> to the telemetry Claude Desktop and Cowork send through a Claude apps gateway, matching terminal sessions</li> <li>Added support for pointing <code>--plugin-dir</code> at a folder of plugi…
<h2>1.4.0 (2026-09-04)</h2> <p>Full Changelog: <a href="https://github.com/anthropics/anthropic-sdk-python/compare/v1.3.0...v1.4.0">v1.3.0...v1.4.0</a></p> <h3>Features</h3> <ul> <li><strong>api:</strong> add Claude Tag category and user breakdowns to usage reports (<a href="http…
<h2>What's changed</h2> <ul> <li>Added an "Organization policy" line to <code>/status</code> and <code>claude doctor</code> that says why your organization's policy could not be loaded, such as a proxy not passing the endpoint through</li> <li>Added <code>bashOutputMaxChars</code…
<h2>Overview</h2> <p>llama.cpp 0.4.0 adds initial Qwen3.8-Flash-Next and Nemotron-3-Puzzle support, on-demand tensor reading, per-slot server context limits, video input options, and a ggml update to 0.23.0 with major sparse flash attention and RDMA work.</p> <h3>API changes</h3>…
<h2>What's changed</h2> <ul> <li>Added a diff panel that opens beside the conversation in fullscreen mode and shows your uncommitted changes as Claude edits; toggle it with <code>/diff</code></li> <li>Added a likely cause for prompt-cache misses (e.g. tool definitions or system p…
<h2>What's changed</h2> <ul> <li>Added <code>managedMcpServers</code> managed setting: organizations can provide HTTP/SSE MCP servers to every user (same entry shape as <code>.mcp.json</code>); entries that name a command to run are skipped</li> <li>Added <code>--permission-promp…
<h2>What's Changed</h2> <ul> <li>Honor GGUF model defined default parameters</li> <li>MLX, MLX-C, llama.cpp update</li> </ul> <h2>New Contributors</h2> <ul> <li><a class="user-mention notranslate" href="https://github.com/marcelpetrick">@marcelpetrick</a> made their first contrib…
<h2>What's changed</h2> <ul> <li>Fixed Claude Code failing to launch on macOS 12 (Monterey), a regression introduced in 2.1.255</li> <li>Fixed remote and scheduled sessions failing with "user messages must have non-empty content" after a re-sent permission approval could not be a…
<h2>What's changed</h2> <ul> <li>Added Claude Fable 5.1 (<code>claude-fable-5-1</code>), now the default Fable model — 1M context, $10/$50 per Mtok with $0.25/Mtok cache reads</li> <li>Added "Time format" (<code>timeFormat</code>) and <code>timeZone</code> settings: 12-hour, 24-h…
<h2>What's changed</h2> <ul> <li>Fixed Bash commands failing with "task output swap refused (tasks dir moved or linked)" on some Macs</li> <li>Fixed "always allow" not saving in a project that has no .claude/settings.local.json yet</li> <li>Fixed Remote Control sessions hosted by…
Claude Agent SDK (Python) — Releases
TIER_1(SK)·github-actions[bot]·
<h3>New Features</h3> <ul> <li><strong><code>snapshot</code> option for system prompts</strong>: Added a <code>snapshot</code> field to <code>SystemPromptPreset</code> and a new <code>SystemPromptCustom</code> typed dict. When <code>snapshot</code> is True, the session keeps the …
Claude Agent SDK (Python) — Releases
TIER_1(SK)·github-actions[bot]·
<h2><a href="https://github.com/openai/openai-python/compare/v3.16.0...v3.16.1">3.16.1</a> (2026-09-18)</h2> <h3>Bug Fixes</h3> <ul> <li><strong>api:</strong> avoid loading unrelated API resources on first use (<a href="https://github.com/openai/openai-python/issues/3898">#3898</…
<h2>What's Changed</h2> <ul> <li>Add new model blueprints and reorganize subgraph categories by <a class="user-mention notranslate" href="https://github.com/comfyui-wiki">@comfyui-wiki</a> in <a class="issue-link js-issue-link" href="https://github.com/Comfy-Org/ComfyUI/pull/1478…
<h2><a href="https://github.com/openai/openai-python/compare/v3.7.0...v3.8.0">3.8.0</a> (2026-09-03)</h2> <h3>Features</h3> <ul> <li><strong>api:</strong> add gpt-6-astra and related features (<a href="https://github.com/openai/openai-python/issues/3791">#3791</a>) (<a href="http…
