Gemini 4 Argon is now available for developers.
What to do: Check the Gemini API Developer Guide for setup details if you want to build with the new model on the paid tier.
AIWire: AI model shutdowns, prices and API changes
The wire
Every change we detect, newest first. Times are Eastern (ET). Archive
TuesdayOctober 13, 2026
Degraded performance in GPT-6.1 Sol Codex and Work mode
Heads-upOpenAI
What to do: No action; resolved.
MondayOctober 12, 2026
Google Gen AI Python SDK 2.30.0
What to do: Update to v2.30.0 if you need streaming continuation resumption or billable token counts; note automatic continuation resumption is disabled by default for generateContent.
ReplaySSM: Rethinking State Storage for Faster Hybrid-Model Decoding
What to do: If you serve hybrid SSM-attention models, try ReplaySSM via the integrated TensorRT-LLM, vLLM, SGLang, or FlashInfer builds and the GitHub code for faster decoding and speculative throughput.
SundayOctober 11, 2026
OpenAI Python SDK 3.29.0
What to do: Update the OpenAI Python SDK to v3.29.0 to pick up the new error codes, Live voice, and vault features.
SaturdayOctober 10, 2026
claude-haiku-4-5-20251001 no longer appears on Anthropic's models page
What to do: Check the model still answers; it is no longer on Anthropic's model list.
FridayOctober 9, 2026
gemini-3.8-flash-tts: Prompt caching supported → not supported
What to do: Check requests that use Prompt caching with gemini-3.8-flash-tts.
gemini-3.8-flash-lite-tts: Prompt caching supported → not supported
What to do: Check requests that use Prompt caching with gemini-3.8-flash-lite-tts.
OpenAI Python SDK 3.28.0
What to do: Update the OpenAI Python SDK to 3.28.0 if you need agent environment suspension and expiration support.
You can now let a Claude Managed Agents agent use dynamic workflows, in beta with the managed-agents-2026-04-01 beta header.
What to do: Add the managed-agents-2026-04-01 beta header and enable the multiagent workflows field to try dynamic workflows, and monitor runs through workflow_run.* events.
Anthropic Python SDK 1.13.0
What to do: Update the Anthropic Python SDK to v1.13.0 if you need the new analytics types or Managed Agents filtering features.
OpenAI Python SDK 3.27.0
What to do: Update the openai Python package to v3.27.0 if you want to use prewarmed hosted environments.
Mistral Python SDK 3.2.0
What to do: Update to mistralai v3.2.0 on PyPI and fix any code using the removed pipeline-config beta operations or old definition variants.
ThursdayOctober 8, 2026
Added Ultrafast mode for GPT-6.1 Sol in the Responses API.
What to do: If you need lower inter-token latency, try service_tier: "ultrafast" with gpt-6.1-sol and review the Ultrafast pricing.
Gemini 3.7 Flash deprecation : gemini-3.7-flash is deprecated and has been replaced by gemini-3.8-flash.
What to do: Update your model strings to gemini-3.8-flash or gemini-3.6-flash, and migrate Deep Research requests to a newer agent version before October 23, 2026.
Deep Research Agent deep-research-pro-preview-12-2025 deprecation : The deep-research-pro-preview-12-2025 agent is deprecated and will be shut down on October…
What to do: Update the agent parameter in your interactions.create requests to deep-research-preview-04-2026 or deep-research-max-preview-04-2026 before the October 23, 2026 shutdown.
WednesdayOctober 7, 2026
Anthropic cut claude-sonnet-5-5 price: cached input $0.20 → $0.10 per 1M tokens
What to do: No code change needed; update your cost estimates.
Google Gen AI Python SDK 2.29.0
What to do: Update to v2.29.0 if you need the new network config option or refreshed sample models.
Supercharge your development with the Google Developer Knowledge API ecosystem
What to do: Try the Developer Knowledge API via the gcloud CLI quickstart, agent skill installer, or client libraries to ground your agents in fresh Google documentation.
In Claude Managed Agents, a cloud environment with limited networking now also applies its allowed_hosts to the web_search and web_fetch tools.
What to do: If you use limited networking with web tools, add any hosts in allowed_domains to allowed_hosts (or remove those entries) to avoid 400 errors on session creation.
The Python and TypeScript SDKs now include classes, in beta, for the browser use tool and the computer use tool.
What to do: If you build browser or computer use integrations, try the new beta SDK toolset classes instead of wiring the tool loop yourself.
The Compliance API chat endpoints now also return chats from the unified Claude experience, in beta for Claude Enterprise organizations, with your existing Com…
What to do: If you use the Compliance API for Claude Enterprise, review the updated chat retrieval docs to include unified Claude experience chats.
Claude Haiku 5.5
New modelReleasesHacker News · community
What to do: Review the migration guide and updated pricing, and consider switching high-volume workloads to claude-haiku-5-5 and updating SDKs for the new computer/browser use features.
Updated the chat-latest snapshot, which points to the latest model available in ChatGPT for Plus, Pro, Business and Enterprise users.
What to do: Use the GPT-6 model family for production workloads and reserve chat-latest for testing, since its snapshot changes regularly.
Claude Max and Team plans now include monthly API credits.
What to do: If you're on a Claude Max or Team plan, claim your monthly API credits via the documented process before assuming you need separate API billing.