Open Source LLM

Gemini 3.1 Pro 正式推出　推理效能較上代翻倍

UNWIRE.HK | Feb. 20, 2026, 8 a.m. | Original

Google 在 2 月 19 日正式推出 Gemini 3.1 Pro 模型。這模型在上週 Gemini 3 Deep Think 更新基礎上進一步強化核心推理能力，在 16 項基準測試中多達 13 項位居榜首並即日向消費者、開發者及企業全面開放。

推理效能大幅提升

3.1 Pro 在測試邏輯

Continue reading...
Open Source LLM

Google 發布 Gemini 3.1 Pro 模型 性能再翻倍 完全壓倒 ChatGPT 5.2

HKEPC | Feb. 20, 2026, 2:46 a.m. | Original
【AI 大戰 ... 😱】距離 Gemini 3 Pro 發布僅三個月，Google 推出全新 Gemini 3.1 Pro AI 模型，提供更強的推理能力、更佳的編碼效能，以及更完善的長文件處理功能，並在大多數基準測試壓倒 OpenAI 最新的 ChatGPT 5.2 AI 模型。
Open Source LLM

NVIDIAから日本語LLM登場。軽量級で最高性能をうたう

pc watch | Feb. 20, 2026, 12:25 a.m. | Original
NVIDIAは、90億パラメータの言語モデル「NVIDIA Nemotron-Nano-9B-v2-Japanese」を公開した。商用利用も可能だとしている。
Open Source LLM

Intel's Discontinued Open-Source OpenPGL Project Finds A New Home

Phoronix | Feb. 18, 2026, 12:33 p.m. | Original
Back in 2022 Intel announced OpenPGL as an open-source library for path guiding to help enhance the quality of path-based renderers. With time Blender began making use of OpenPGL and other industry interest and adoption. Unfortunately, Intel quietly ended work on OpenPGL in 2025 but has now fortunately found a new home.

The Apache 2.0 licensed Intel Open Path Guiding Library was started as part of the Intel (oneAPI) Rendering Toolkit and found use by not only Blender but also OpenMoonRay as
Continue reading...
Open Source LLM

I started using my local LLM with Obsidian and should have done it sooner

XDA developers | Feb. 18, 2026, 10:30 a.m. | Original
Obsidian is already great, but my local LLM makes it better
Open Source LLM

Claude Sonnet 4.6 正式推出　評測勝 Opus 4.6 成 Free 及 Pro 用戶預設模型

UNWIRE.HK | Feb. 18, 2026, 6 a.m. | Original

Anthropic 於 2 月 17 日正式推出 Claude Sonnet 4.6，並將其設定為 Free 及 Pro 方案用戶在 claude.ai 與 Claude Cowork 應用程式中的預設模型。新版本在編程、電腦操控、長文本推理及知識型工作等範疇均有顯著提升，API 定價則與上代 Sonnet 4.5 維持一致。

性能突

Continue reading...
Open Source LLM

NotebookLM、パワポ出力に対応。Googleスライド形式も近日中に

pc watch | Feb. 18, 2026, 1:56 a.m. | Original
Googleが提供するAIリサーチツール「NotebookLM」に、PPTX形式ファイルの出力機能を実装したと、公式X(旧Twitter)アカウントの投稿で発表された。
Open Source LLM

Claude Sonnet 4.6 model brings ‘much-improved coding skills’ and upgraded free tier

9to5mac | Feb. 17, 2026, 1:33 p.m. | Original
Claude Mac app

Anthropic just released the second Claude model upgrade this month. Claude Sonnet 4.6 is the first upgrade to Anthropic’s medium-sized AI model since version 4.5 arrived in September 2025.

Anthropic says Claude Sonnet 4.6, which features a “1M token context window,” delivers a “full upgrade of the model’s skills across coding, computer use, long-context reasoning, agent planning, knowledge work, and design.”

Sonnet 4.6 brings much-

Continue reading...
Open Source LLM

人間と同じようにPCを操作できるAI、Sonnet 4.6登場

pc watch | Feb. 17, 2026, 1 p.m. | Original
Anthropicは18日(日本時間)、Sonnetモデルとしては過去最高性能を持つ「Claude Sonnet 4.6」を発表した。Claude.aiにおける新たなデフォルトモデルとして採用され、Claude CodeおよびCoworkのすべてのプランで利用可能。
Open Source LLM|Intel

(PR) AMD Announces Day 0 Support for Qwen 3.5 LLM on Instinct GPUs

TechPowerUp News | Feb. 17, 2026, 9:24 a.m. | Original
AMD is excited to announce Day 0 support for Alibaba's latest generation of Large Language Models, Qwen 3.5, on AMD Instinct MI300X, MI325X, and MI355XGPU accelerators, in close collaboration with Alibaba Qwen team. Leveraging the optimized ROCm software stack, SGLang and vLLM inference serving framework, developers can immediately deploy these state-of-the-art models.

The Strategic Value of Qwen 3.5 on AMD Instinct
Empowering Next-Generation AI
Continue reading...
Menu