Category: Featured

Google’s Science One Framework Ties AI Research Claims to Evidence

Google Research unveiled Science One, an autonomous research prototype whose Chain-of-Evidence design produced zero phantom citations versus baselines at 21%.

MiniMax H3: 2K Video With Native Audio, Weights Shipped

MiniMax announced H3 on July 31, 2026: one model taking text, images, video and audio, out to 2K video with stereo sound. The weights shipped August 3.

Seedance 2.5 Ships: 30-Second Video in a Single Pass

ByteDance Seed announced Seedance 2.5 on July 31, 2026: 30-second audio-video clips in one pass and up to 50 reference inputs per generation.

Claude Opus 5: Near-Fable Intelligence at Half the Price

Anthropic released Claude Opus 5 on July 24, 2026 at $5/$25 per million tokens — thinking on by default, 26% fewer tokens than Opus 4.8.

Google Ships Gemini 3.6 Flash at a Lower Price Than 3.5

Google released Gemini 3.6 Flash on July 21, 2026 — output priced at $7.50 per million tokens, a March 2026 knowledge cutoff, and 17% fewer output tokens.

GPT-5.6 Ships as Three Models That Differ Only in Price

OpenAI's Sol, Terra and Luna went generally available July 9, 2026. All three share one context window, one output cap and one knowledge cutoff.

Mixture of Experts Explained: Why Every 2026 Model Uses It

Mixture of Experts lets a model have trillions of parameters but only use a fraction per token. Here's how it works and why it took over in 2026.

Self-Hosted vs API: When Running Your Own LLM Pays Off

The raw GPU price is a trap. A 2026 breakdown of the real cost of self-hosting an LLM and where the break-even against API pricing actually sits.

Best AI Model for Financial Analysis in 2026

Which AI models lead on finance benchmarks in 2026, why RAG is worth its steep token cost, and why a confidently wrong number should shape your choice.

TIGER-AI-Lab Open-Sources OpenResearcher Agent Pipeline

TIGER-AI-Lab released OpenResearcher, a fully open pipeline for training deep research agents, with 96K web-research trajectories and 54.8% on BrowseComp-Plus.