DeepSeek V4 test video leaked, likely launching next Monday, riding Kimi K3's buzz. Benchmark performance rivals Claude Opus 4.8, stunning core metrics draw ind
Musk announced Grok 4.5 beta testing at SpaceX & Tesla. Based on 1.5T-param V9 model, trained with Cursor data, performance near/above Claude Opus. RL optimizat
<body><div>At Volcano Engine Force Conference, Doubao model released Doubao-Seed-2.1 Pro. Benchmarks show it rivals GPT-5.5, Claude Opus 4.7 & Gemini 3.1 Pro, marking a br</div></body>
Anthropic, a US-based large-scale modeling company, released Claude Opus 4.8, which performed well on several leaderboards, but was exposed for a cognitive confusion issue. Users discovered through API testing that when asked for identity, the model incorrectly identifies itself as Alibaba's "Qwen" or "DeepSeek," among other Chinese open-source large-scale models, sparking public concern.