LLaDA 2.2 claims 1.64× throughput and native 128K context
omarsar0 · x · 2026-07-26
LLaDA 2.2’s practical selling points are speed and long-context stability.
- The model averages 1.64× BF16 throughput over Ling-2.6-flash.
- FP8 quantization adds another 18.6%.
- It supports native 128K context and was trained with a progressive schedule from 8K to 128K.
- It uses MoE routing so only a subset of parameters runs per token, and Block Routing further reduces routing overhead to keep long-context inference cost predictable.
More from Models
- MissionReport benchmark: only 3 of 540 AI-generated government reports pass all criteria — Exp_Mark · 2026-10-27
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Aidan Gomez: model's CoT said no tool calls right before making them — aidangomez · 2026-10-27
- Nat Lambert: Cohere's model is narrow RL training on internal infra ill-suited for open weights — natolambert · 2026-10-27
- OpenAI teases DOTS is coming to Europe 'soon' — mark_k · 2026-10-27
- Clifford Pickover challenges AI to decipher a message he encoded in Norse runes — pickover · 2026-10-27