Releasesby IbisTypeScript v2 and Python functions now support Global Branching
GA11 days agoUnverifiedAdded Oct 18, 2026
Global branching is generally available for TypeScript v2 and Python functions in Foundry. Developers can import branched Ontology entities into code, test changes on branched data with live preview, and publish branch-isolated pre-releases for use in downstream workflows.
Topics: Developer tools
More Foundry and AIP releases
Every Foundry and AIP releaseControl LLM data access with markings in AIP
Administrators can configure an allowlist of approved mandatory markings in Control Panel to restrict LLM data access. AIP blocks any LLM request from a Foundry session containing access to mandatory markings outside the allowlist.
GLM-5.3 now available in AIP through Amazon Bedrock
Z.ai's GLM-5.3 open-weight reasoning model is now available in AIP as an experimental model through Amazon Bedrock for non-georestricted and US-georestricted commercial enrollments. It supports tool calling, structured outputs, and streaming responses for software engineering and agent tasks.
Claude Haiku 5.5 now available in AIP
Anthropic's Claude Haiku 5.5 is available in AIP for eligible commercial enrollments and US government IL2 and IL4 enrollments. It is designed for fast responses and high-volume workloads such as summarization, classification, and focused coding tasks.
GPT-6.1 Sol now available in AIP
OpenAI's GPT-6.1 Sol model is available in AIP through Direct OpenAI, Azure OpenAI, and Amazon Bedrock for eligible commercial enrollments.
Gemini 4 Argon from Google Vertex AI now available in AIP
Gemini 4 Argon from Google Vertex AI is available as an experimental model in AIP. It supports commercial non-georestricted, US georestricted, and EU georestricted enrollments, as well as US Government IL2, IL4, and IL5 enrollments.
View state break causes directly in Pipeline Builder
Pipeline Builder's deploy menu now displays the specific reasons causing state breaks for streaming and incremental pipelines. Users can view detailed change breakdowns, navigate directly to affected graph nodes, and edit or revert changes in place.
Developer Console now supports Global Branching
Developer Console now supports Global Branching, allowing developers to build, test, and review Ontology SDK React applications alongside Ontology changes on the same branch. Capabilities include generating branched SDKs, testing with branch data, and sharing pull request previews.
Foundry Worker operations available without compute costs
Foundry Worker data ingest and export, media ingest, and Iceberg extract operations are available without compute costs. This update aims to facilitate migration from Agent Worker to Foundry Worker. Streaming ingests will continue to be billed separately.
Also shipped on Oct 13, 2026
Palantir in October 2026What's included:
- Three training methods: Supervised fine-tuning (SFT) , direct preference optimization (DPO) , and reinforcement learning with group relative policy optimization (GRPO) . Each method trains a low-rank adapter (LoRA) or the full model weights. Custom loss functions : Train with any objective you compute in PyTorch from the model's log-probabilities.
The AI.AGG function is generally available (GA)
The AI.AGG function is generally available (GA). You can use this function to semantically aggregate unstructured input data based on natural language instructions.
BigQuery generative AI functions now support the gemini-4-argon model
BigQuery generative AI functions now support the gemini-4-argon model. This model is generally available .
Gemini 4 Argon in CodeMender (Preview)
CodeMender now supports Gemini 4 Argon. To use it, pass --model gemini-4-argon to cm find , cm verify , or cm fix . Gemini 3.8 Flash remains the default model. For more information, see Specifying the model .
Gemini 4 Argon and Gemini 4 Argon Cyber Permissive
Gemini 4 Argon ( gemini-4-argon ) is available in General Availability (GA) . Gemini 4 Argon is Google's frontier model that excels at deep work in complex, real-world tasks across domains including coding, finance, legal, cybersecurity, and science.
On-demand state repartitioning is now generally available
On-demand state repartitioning is now generally available. You can resize the number of shuffle partitions for a stateful Structured Streaming query or Lakeflow pipeline without losing checkpoint state. See On-demand state repartitioning for stateful streaming queries .