I tested ox alpha ( Zai_org glm-5.3-flash) vs AnthropicAI Opus 5 on threejseval!
We just added it yesterday before the big reveal. It is an incredibly good model, now rank second just before Opus 5 AND it costs around 400 times less! immediatly reached the forntier.
Claude Opus 5 vs Gemini 3.7 Flash for computer use.
Open and create a short visual song called RAINBOW RUSH. On the melody grid, draw one clear… Claude Opus 5 won against Gemini 3.7 Flash. Go try computer-use agents for free at
Claude Opus 5 vs GPT-5.6 Sol for computer use.
Open and use only the visible AutoDraw, shape, fill, and text tools to create a bold poster called… Claude Opus 5 won against GPT-5.6 Sol. Go try computer-use agents for free at
A 23 year old guy in China is running 300 AI agents at the same time
Built a system where 300 Kimi K3 agents investigate 100 EV companies in parallel. Opus 5 then checks your work against the original sources. first round: 12 agents
You have reached the end of the archive
All of Claude Opus 5