Anthropic shipped @anthropic-ai/[email protected] on 2026-09-28 (36 commits ahead of 0.128.0): claude-sonnet-5-5, between_tools thinking, cache diagnostics GA, Managed Agents event type filters, ClientToolUnion, and optionally starting tool calls while the reply streams—plus streaming tokenizer/partial tool-input fixes. Mirrors the Python 1.9.0 surface for TS agents.
Key Takeaways
- ✓Release sdk-v0.129.0; 36 commits vs 0.128.0
- ✓Adds
claude-sonnet-5-5model id; models docs https://docs.anthropic.com/en/docs/about-claude/models - ✓
between_toolsthinking + optional tool starts during reply streaming;ClientToolUnion - ✓Cache diagnostics GA; typed Managed Agents event filters; rate-limit
include_inherited/source - ✓
npm i @anthropic-ai/[email protected]; https://www.npmjs.com/package/@anthropic-ai/sdk ; https://docs.anthropic.com/en/api/messages
Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Background
TS agents needed parity with Python 1.9.0 for Sonnet 5.5, between_tools thinking, and streaming tool starts. @anthropic-ai/[email protected] closes that gap.
What shipped
sdk-v0.129.0: claude-sonnet-5-5, between_tools, ClientToolUnion, cache diagnostics GA, Managed Agents typed filters, optional tool starts during streaming; tokenizer/SSE perf fixes. 36 commits ahead of 0.128.0.
Benchmarks
No SWE-bench numbers in the notes. Anchors: ahead_by 36; publish 2026-09-28T18:03:48Z. Measure your own between_tools latency.
Get started
npm i @anthropic-ai/[email protected]. Docs: https://docs.anthropic.com/en/api/messages
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.