AI.info
AI Tools Updates — Page 67
What changed in the AI tools this week — dated, from the makers’ own changelogs.
En anglais
- Kilo Code — v7.8.12 (release)
Routes gateway traffic through ai-gateway.kilo.ai by default, adds a URL override, and changes ignored-path handling in glob and grep.
- Opik — 2.2.102-7069
- Opik — 2.2.102
Added a model catalog with profiles and per-model params; indexed trace-keyed span reads; restricted Ollie widget building to Custom Charts.
- Opik — 2.2.102-7068
Bound trace-keyed span reads using the span_weeks index.
- Opik — 2.2.102-7067
- Keploy — v3.6.127
Added S12 exact serving and MySQL credential redaction, and fixed the TCP DNS redirect leak.
- Snowflake Cortex Analyst — Data mirroring for Snowflake Postgres (General availability)
Data mirroring for Snowflake Postgres
- llama.cpp — b11658
Added support for the maion-coder architecture and its draft models.
- llama.cpp — b11644: kv-cache: honor PARTIAL_ONLY in the plain KV cache (#30514)
The plain KV cache now honors PARTIAL_ONLY when reading and writing cache state.
- llama.cpp — b11669
llama.cpp version was bumped to 0.7.0, with build artifacts listed for macOS, iOS, Linux, and Android.
- llama.cpp — b11645
Added a BF16 variant of the Metal tensor API flash attention kernel.
- llama.cpp — b11647
Vulkan now copies strided f16 KV data once on AMD so reads spread across memory channels.
- llama.cpp — v0.7.0
v0.7.0 adds LiliCorr speculative decoding, GLM5-Next MTP, a GPU cache for MoE experts, new model support, and a refactored chat session API.
- llama.cpp — b11655
- llama.cpp — b11643
Embedding-only graphs now expand t_embd when embeddings are off so the graph is not empty.
- Palantir AIP — View state break causes directly in Pipeline Builder
The deploy menu surfaces the reasons for unacknowledged state breaks for streaming and incremental pipelines.
- Palantir AIP — Claude Haiku 5.5 now available in AIP
Anthropic’s Claude Haiku 5.5 is now available in AIP for eligible commercial enrollments and US government enrollments at IL2 and IL4.
- Palantir AIP — Developer Console now supports Global Branching
Develop, test, and review Ontology SDK (OSDK) React application changes alongside Ontology changes on the same global branch.
- Palantir AIP — Gemini 4 Argon from Google Vertex AI now available in AIP
Gemini 4 Argon from Google Vertex AI is now available as an experimental model in AIP.
- Palantir AIP — GPT-6.1 Sol now available in AIP
OpenAI’s GPT-6.1 Sol is now available in AIP through Direct OpenAI, Azure OpenAI, and Amazon Bedrock for eligible commercial enrollments.
- Palantir AIP — GLM-5.3 now available in AIP through Amazon Bedrock
GLM-5.3 is available as an experimental model in AIP through Amazon Bedrock.
- Everlaw Ediscovery Platform — Everlaw and Microsoft Copilot Announce New MCP Integration (Available Now)
A new MCP integration brings Everlaw’s evidentiary system of record to Microsoft 365 through Copilot.
- Everlaw Ediscovery Platform — Everlaw and Thomson Reuters CoCounsel Legal Launch New Partnership
Everlaw and CoCounsel Legal announced a forthcoming integration connecting matter facts with the law in AI tools.
- Everlaw Ediscovery Platform — Everlaw Brings Frontier AI Closer to the Evidence with ChatGPT Plugin
A new plugin connects Everlaw’s governed evidence layer with ChatGPT Enterprise.
- Everlaw Ediscovery Platform — Everlaw for Word Brings Your Evidence Directly to Where You Draft
Everlaw for Word brings evidence into Word, lets users ask questions across deposition transcripts, and manage designations in the transcript.
- Krea AI — Krea 3 API in early access, with Zero Data Retention and renamed thinking modes
Krea 3 API supports image editing, text-to-image, up to 8 reference images, regional edits, and 1K or 2K output; ZDR is available.
- RunPod Serverless — Breaking runpod 2.0 CLI changes
runpod 2.0 removes the runpod config, runpod project, and runpod exec commands. Use rp login and rp flash commands.
- RunPod Serverless — Deprecation Standalone Flash package
The runpod-flash package no longer gets new features, and the runpod/flash repository will be archived.
- RunPod Serverless — New Release Runpod SDK 2.0
Flash is now part of the runpod Python package; define queue endpoints, load-balanced APIs, and tasks that run on their own Pod.
- n8n — n8n@2.43.6
Create an XML parser for each webhook request body.
- n8n — n8n@2.44.2
Create an XML parser for each webhook request body.
- n8n — stable
Create an XML parser for each webhook request body.
- Hugging Face — Apertus 1.5 Release Notes
Released Apertus 1.5.
- Figma AI — Here’s everything from Config India 2026
Launched new ways to build and collaborate with code, capture and share context, and build with real data and content.
- Zed — Zed v1.24.2
Added Git tag creation, Terminal tab dragging into the Agent Panel, and File Finder search prefill from selected text.
- Rox — Contact cards in Sequences
Hover over a contact’s name to see their details, opt-out status, last activity, and other sequences they’re enrolled in.
- XO — XLN Online Installer
Fixed issue with locking up on first install
- Truss — v0.18.35
Set sampler replica count in truss loops push and labeled startup logs until readiness.
- Truss — v0.18.36rc0
Fixed Chains reference generation and added batch System One API.
- Mistral Vibe — Workflow Visualizer
Studio renders any Workflow as a read-only graph of its activities and execution paths, generated automatically from your code.