Open release: Latest AI News

Perplexity Open-Sources Lily, a Qwen-Specific Mac Engine That Beats MLX-LM in Its Test

The Rust-and-Metal runtime gives developers a public, deployable version of Perplexity’s local inference stack. Its reported gains come from one Qwen model, one hardware family and a batch-one comparison on a high-memory M5 Max.

Sep 1, 2026Tools

AWS Publishes Jamf’s Bedrock Budget Gate, Restricting Premium Models at 80% Spend

The open sample makes AI cost controls more operational: developers can keep working on a cheaper model, while premium access is automatically narrowed as each person approaches a daily limit.

4 min

Last week

9 stories

Aug 28, 2026Models

Gnani Launches 30B Evon Model and Self-Hosted Agent Stack for Indian Institutions

The open-weight release couples Indian-language support with tools intended to keep sensitive data inside a customer’s own infrastructure.

Aug 28, 2026Infrastructure

Lightwheel Releases 10,000 Hours of Human Video for Robots, With 90,000 Still to Come

The release gives robotics teams a sizable pretraining resource and a way to test two collection designs, but its value will depend on annotation quality, task fit, and robot-specific data added afterward.

Aug 26, 2026Models

Z.ai Releases 320B GLM-5.3-Flash With MIT Weights, 1M Context and Low API Rates

The model gives developers a permissively licensed route to long-context, image and video workloads, while its performance and serving-efficiency claims remain company-reported results.

2 min

Aug 26, 2026Models

Aug 25, 2026Infrastructure

Perceptron Releases Open-Weight Isaac 0.5 to Span Warehouse Robot Tasks

The startup’s pitch turns on a difficult handoff: translating visual understanding into sequenced work on a factory floor.

2 min

AWS Publishes a Phone-Ordering AI Pattern That Connects Restaurant Data to Claude

The sample shows how a voice agent can take an order from a phone call through backend tools, but restaurants must supply and operate the systems holding their own menus, carts, orders and locations.

Aug 25, 2026Tools

Liquid AI’s Pipette Tests 1,000+ On-Device AI Setups—and Limits Cross-Device Rankings

The open-source suite puts model, quantization, runtime and hardware in one result. Its quality scores still come from H100 reference systems, while early results are not designed for cross-device comparisons.

Aug 25, 2026Tools

Aug 24, 2026Infrastructure

Nvidia’s NeMo Switchyard Routes Agent Calls Across Models, Not One Default

The open-source routing layer can switch models as an agent encounters tool results and errors, but companies must now test the routing policy itself as prices and capabilities move.

4 min

Meta Sends MetaRoCE to OCP for Loss-Tolerant AI Ethernet

The proposed RDMA transport treats out-of-order arrival and packet loss as normal network conditions, shifting more control to endpoint NICs. Meta’s reported results are promising, but hardware interoperability remains the consequential test.

Aug 24, 2026Tools

AWS Publishes Metadata Workflow That Escalates Ambiguous Fixes to Bedrock LLMs

The deployable sample uses validation and similarity methods before LLM field resolution. AWS documents both a human-approved workflow and an agent that can apply corrections autonomously, while leaving operating benchmarks for adopters to establish.

2 min

August 2026

3 stories

Aug 22, 2026Tools

TrueForge Puts the Agent Harness, Not the Model, at the Center of Cost Control

The open-source release turns the session layer around an AI model into a deployable product. Its reported savings are promising, but rest on a small enterprise benchmark and leave real-world transfer unproven.

Aug 22, 2026Models

Thinking Machines Releases Inkling, but Argues Open Weights Need a Gate

The lab’s argument is not that every capable model should be closed. It is that release decisions should turn on whether a model adds risk beyond what is already downloadable—and whether defenders have had time to prepare.

Aug 19, 2026Tools