Alibaba’s Qwen 3.6-35B-A3B is a sparse MoE powerhouse with 3B active parameters, a 1M token context, and a new ‘thinking preservation’ mode for complex agentic workflows.
Read More
Tag: benchmarks
Claude Opus 4.7: 1M Context and the 128K Output Frontier
Anthropic drops Opus 4.7 with a massive 1M token context window, 80.9% SWE-bench score, and a new tokenizer that boosts efficiency by 35%.
Read More