https://llmrequirements.com/news/2026-06-30-claude-sonnet-5-vs-glm-5-2 LLMRequirements.com en 2026-06-30T00:00:00.000Z Claude Sonnet 5 vs GLM 5.2: the closed benchmark king vs the open model you can actually run software, models, comparison, anthropic, open-weights https://llmrequirements.com/news/2026-06-30-dgx-spark-122b-dflash-speculative-decoding LLMRequirements.com en 2026-06-30T00:00:00.000Z DGX Spark runs Qwen3.5-122B at 59 tok/s general and 81 on agent traffic, with speculative decoding (not NVFP4) performance, nvidia, dgx-spark, speculative-decoding, qwen, agentic