Technology ❯ Artificial Intelligence

Model Performance

Benchmarking Comparative Analysis Evaluation Metrics Benchmarks Reasoning Capabilities Parameter Count User Experience Cost Efficiency User Feedback Cost Analysis

NVIDIA Releases Nemotron 3.5 Lightning and Open Router to Speed Cheap, Local Agent Workloads

NVIDIA says the tools will cut runtime costs for multi‑model agents by routing routine steps to faster, smaller models.

Alibaba Publishes Qwen 3.8–Max and Will Release Open Weights Next Week

Z. AI Releases GLM-5.2 With 1 Million‑Token Context and Open MIT Weights

Google Brings Gemma Models and On‑Device AI Tools to macOS

DeepSeek Debuts V4 Preview With 1M‑Token Context, Tuned for Huawei Chips

NVIDIA Confirms Blackwell Support for DeepSeek‑V4 With 150 Tokens a Second

Xiaomi Debuts MiMo‑V2.5 AI Family, Unifying Text, Image, Audio and Video

Google DeepMind Releases Gemma 4, Open AI Models Built for Phones to Workstations

OpenAI Releases GPT‑5.4 Mini and Nano for Low‑Latency, Low‑Cost AI Tasks

OpenAI Ends GPT-4o in ChatGPT, Prompting Backlash and Copycat Services

Google Releases T5Gemma 2, a Compact Multimodal Encoder-Decoder With 128K Context

Nvidia Launches Open-Source Nemotron 3, a Three-Tier MoE Model Family for Multi-Agent AI

OpenAI Said to Be Building 'Garlic' to Challenge Gemini 3 and Opus 4.5

DeepSeek Releases V3.2 Models to Challenge GPT‑5 and Gemini 3 Pro

OpenAI Starts Rolling Out GPT-5.1 and Pilots ChatGPT Group Chats

PaddlePaddle Releases PaddleOCR-VL, a 0.9B Vision-Language Model for Multilingual Document Parsing

OpenAI Says Incentives Drive AI Hallucinations, Calls for Scoreboard Overhaul

OpenAI and Anthropic Publish Reciprocal Safety Tests of Each Other’s AI Models

Z.ai Unveils Open-Source GLM-4.5 Models That Undercut DeepSeek

Google’s Gemini 2.5 Pro and Flash Go Live as Cost-Efficient Flash-Lite Debuts in Preview