We put Zai_org's GLM-5.3 up against Kimi K3, DeepSeek V4 Pro, and Qwen 3.8-Max in Flotilla, our naval resource-and-diplomacy sim.
It won the 5-game series 3-2, and was the most consistent at relaying live orders to ships already at sea. 🗺️
PERPLEXITY 🔥: A Hybrid mode for Perplexity Computer on Mac has been officially announced!
The Hybrid mode is powered by PPLX Qwen 3.8 27B, a custom post-trained model from Perplexity. It also comes with a Privacy Gate feature to detect PII data before it is sent to the
Tested GLM 5.3 Flash and Qwen 3.8 Max against the best frontier models using the exact same Three.js + HTML prompt.
The prompt: Recreate an interactive Kage Temple experience with Three.js - animations, interactions, and a full-page 3D environment, all packed into a single HTML
Perplexity just announced Hybrid Compute for Mac.
🔥 It splits workloads between cloud models for research and local models for sensitive data and on-device actions. It runs PPLX Qwen 3.8 27B locally, with one-click setup, 24 GB memory minimum and 32 GB recommended.
You have reached the end of the archive
All of qwen38