Key points in rapid AI progress, A secretive “ox alpha” model (widely suspected to be the latest Chinese GLM open-source model) shows strong coding gains (~80% on a tough benchmark vs. the 60s for Claude/GPT-style models). Nvidia + a new “harness” hit a perfect 100% on ARC-AGI-3. Intelligence-density progress Chinese Qwen-style 27B models already match … Read more
View Source