Z.ai’s open-weights GLM-5.2 beats GPT-5.5 on multiple long-horizon coding benchmarks for 1/6th the cost

It allows engineering teams to host frontier-level AI on their own sovereign infrastructure, entirely eliminating vendor lock ...
techtimes

GLM-5.2 Open Weights Live: Top Coding Benchmark, but API Use Carries China Data Risk

a mobile phone's screen showing the logo of Chinese AI Zhipu in Beijing on January 21, 2026. Investor confidence in Chinese AI startups is riding high, but obstacles to their long-term success range ...

Waterloo's PAW compiles task specs into 23MB LoRA adapters a 600M-parameter model runs entirely offline.

Local AI inference at 32B-parameter quality, no cloud API required: University of Waterloo researchers released PAW on July 2 ...
Tech Times

Compile Once, Run Offline: New AI Method Matches 32B Models With a 23MB File

Local AI inference at 32B-parameter quality, no cloud API required: University of Waterloo researchers released PAW on July 2 ...
Opinion

5 More AI Predictions For The Year 2030

Two years ago, we published a list of 5 predictions about AI in the year 2030. The article sparked a lot of fascinating (and sometimes heated!) discussion. Today, we’re running it back with five more ...
IEEE

Resolving Range Ambiguity via Multiple-Input Multiple-Output Radar With Element-Pulse Coding

In this paper, we propose a novel radar framework, referred to as multiple-input multiple-output (MIMO) radar with element-pulse coding (EPC). The EPC employs Fourier basis in joint transmit elements ...
IEEE

Lattice coding and decoding achieve the optimal diversity-multiplexing tradeoff of MIMO channels

Abstract: This paper considers communication over coherent multiple-input multiple-output (MIMO) flat-fading channels where the channel is only known at the receiver. For this setting, we introduce ...
note

GPT-5.6 Thorough Verification: Decoding the Capabilities of Sol, Terra, and Luna Through Benchmarks and Primary Information

OpenAI's new model, "GPT-5.6" (Sol, Terra, Luna), began a limited preview on June 26, 2026. Following an unusual public release process that started with approximately 20 companies at the request of ...
note

Laguna XS 2.1 Overview: Local Agent Coding Model Enhances Multilingual Performance and Terminal Tasks

While it retains the same "33B-A3B" configuration as the previous Laguna XS.2 model, it is characterized by performance boosts through training and tuning. For comparison, by listing Qwen3.6 (35B-A3B) ...