List your company

Planning on Countdown (Score and Inference Speedup)

15.3Score (%)Full
14.57214.76114.9515.139Mar 5, 2026

Evaluation results

MethodLinks
FullModel=LLaDA-8B, Inference Strategy=Full2026.0315.3———
LSPModel=LLaDA-8B, Inference Strategy=LSP2026.0315.32.63——
LSPModel=Dream-7B, Inference Strategy=LSP2026.0315.12.4——
FullModel=Dream-7B, Inference Strategy=Full2026.0314.6———
Full-step decodingModel=LLaDA-8B-Instruct, Decoding Strategy=Full-step, Evaluation Setting=8-shot2025.08——15.3—
Full-step decodingModel=Dream-7B-Instruct, Decoding Strategy=Full-step, Evaluation Setting=8-shot2025.08——14.6—
ProphetModel=LLaDA-8B-Instruct, Decoding Strategy=Prophet, Evaluation Setting=8-shot2025.08—2.6715.30
ProphetModel=Dream-7B-Instruct, Decoding Strategy=Prophet, Evaluation Setting=8-shot2025.08—2.3714.60