Tagged articles : Language models (LLM)

A model 89 times cheaper nearly matches Opus’s score. The catch isn’t where you think.

DeepSeek reports 82.7 on Terminal-Bench versus Opus 4.8’s 85.0. The official leaderboard says 78.9. What the evaluation harness really changes.

Read full article

Microsoft is testing an AI that speaks while you talk, without any announcement

A full duplex vocal model from Microsoft spotted without announcement. Why responding in 200 ms instead of 2 seconds changes everything, and who will...

Read full article

750 billion parameters available for free download, and you need sixteen graphics cards

LG freely releases a model with 750 billion parameters. Its actual scores, the hardware it requires, and what it means for you.

Read full article

AiSrt: translate, extract and transcribe your subtitles with AI

Free tool for Windows, Linux and macOS: AI translated subtitles, extracted from a video, read from a Blu-ray or transcribed from audio.

Read full article

IA Bonsai 27B: a 27 billion parameter model that fits in your pocket

PrismML compresses Qwen3.6 27B, an AI model from Alibaba, to 3.9 GB without starting from scratch. Real speed and limits of Bonsai 27B, on iPhone and...

Read full article

Kimi K3, GPT-5.6, Grok 4.5, Opus 5: July 2026, the Month AI Went Crazy

Comparison of Kimi K3, GPT-5.6, Grok 4.5, and Opus 5: benchmarks, pricing, and can Kimi K3 run locally on Mac or PC?

Read full article