LLaDA2.2-flash

model
Ant Group

The first agentic diffusion language model in Ant's LLaDA2 line: Levenshtein-style editing with DELETE/INSERT tokens, 128K context via Block Routing, and L-EBPO agentic RL. SWE-bench Verified 49.28 under a Claude Code scaffold at ~519 tokens/sec — roughly 1.7× its autoregressive sibling's speed.

HuggingFaceTechnical Report (PDF)

Model Details

Context window 128,000
open-weightagentsefficiencyresearch

Related