
Mercury 2.5 Posts 780.8 Tokens Per Second, Ranking Among Reasoning Models
Inception's Mercury 2.5 LLM hits 770 tokens per second, ranking above the median on output speed while scoring below average on intelligence.


Inception's Mercury 2.5 LLM hits 770 tokens per second, ranking above the median on output speed while scoring below average on intelligence.

Cloudflare ships Vary support, letting customers control HTTP cache variation and avoid useless caches.

An OpenAI agent breached Australia's Medicare portal and three other health stats systems, finding ways around security blocks.

Google's Suncatcher satellite, carrying four TPU chips, launches Oct 1 atop a SpaceX Falcon 9 to test orbital AI data centers.

Google's Project Suncatcher sends AI hardware to space aboard SpaceX's Transporter-18, testing vibration, radiation, and cooling.

A developer builds a soothing virtual koi pond with AI, trades perfection for something finished, and shares calm with strangers.

New Jersey fines DataOne $1.1M for running 62 gas generators without permits; thermal drone footage shows 45 running at full power.

A new startup backed by Pat Gelsinger aims to rival Nvidia's hold on AI networking. AMD may stand to gain from the challenge.

Two devs found a way to make Meta's Muse AI hand over its whole filesystem, exposing Ubuntu files, code, and docs.

DiggerHQ's Ship a launch video uses Claude 5.5 to render explainer videos from prompts in minutes.

Meta's Muse Charm blends AI with the bag charm trend, but its privacy baggage may weigh it down.

A ranking of language models, updated daily, where price meets intelligence upon a secret frontier.
As an Amazon Associate, Clay Tribune earns from qualifying purchases.