Artificial Analysis's Intelligence Index v5 relies on secret test data. Open A.I. models lose up to 16 points, while Claude ...
SoMark released OXR-1.0 on October 20, 2026, an open-source 1.3B-parameter model that turns PDFs and document images into ...
Liquid AI LFM3 ships 6 open-weight 2B-class models; LFM3-1.7B averages 49.37 across 29 self-reported benchmarks, above ...
The CDC’s latest available measured data puts the average weight of U.S. men age 20+ at 199.0 pounds and women at 171.8 ...
Qwen Image 3.0 ended Alibaba's open-weight image line. Microsoft's 2.7 has a score and no endpoint. Two closed models, two ...
Warm up with these before your next bench day.
If you think heavy bench pressing is something you can only enjoy until a certain age, think again. Shoulder pain is ...
Moonshot AI introduces Kimi K3.1 Pro. According to independent benchmarks, the open-weight model competes with leading ...
Kimi K3.1 Pro ranked fourth among 215 AI models on an independent intelligence index while Moonshot AI deliberately removed ...
Moonshot AI has launched Kimi K3.1 Pro, a new model aimed at coding agents, financial research and 3D creation. The Chinese ...
DeepSeek V4 Pro tops Aikido's cybersecurity benchmark days after Google gated Gemini 4 Argon over the same ...
Moonshot AI has set a deadline for its latest release, but the absence of supporting documentation on its primary ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results