Introducing newsliquid-flash-1.0 Our new specialist model built for real-time market news triage. It ranks #1 on the News Impact Benchmark v1, delivering an industry-leading 13.290 bits/s information rate with just 140.1 ms mean latency.

In financial markets, a correct interpretation that arrives after the market has moved is rarely useful. While public LLM leaderboards typically focus only on accuracy, we evaluate models based on Information Rate—combining per-call quality with throughput to measure bits per second of correct interpretation.

Under a strict single-news, no-batching, no-cache protocol, newsliquid-1.0-flash outperforms frontier models like GPT-5.5 and Claude Opus 4.7 on the combined score. ⚡️ Mean latency: 140.1 ms 🛡️ P99 latency: 226.2 ms 🚀 Throughput: 7.140 QPS

We made a deliberate speed-for-accuracy trade-off. While frontier models hit ~93% accuracy but take around 2 seconds per call, newsliquid-1.0-flash delivers 72.5% output accuracy in just 140ms. It is purpose-built for the operational regime where timeliness is part of the value function.

To help the industry measure what actually matters for trading pipelines, we are also releasing the News Impact Benchmark v1 to the public today. It includes 200 timestamped single-news items across 7 catalyst categories, complete with direction, impact scores, and analyst rationales.

newsliquid-flash-1.0 is already live in production. You can experience its real-time market judgments today—now fully integrated into the newsliquid triage engine and opennews skills. The benchmark dataset and scoring scripts are also available now. Let's move at the speed of the market.

Woc

看起来挺酷

好像不是开源的，不知道多大，手机上能跑么

？

不知道是否会有订阅制

有点意思

alpha:可以通过反方向插针来 拉爆/打止损 所有机器人😂

holy shit Newsliquid is killing the game!

具体怎么用呢

@grok I didn't understand, please help me search for relevant information and analyze it for me.