# llama.cpp update adds warnings for draft‑p‑min usage

> The new release warns users when the p_min flag is combined with probabilistic draft sampling, preventing unintended token drops.

Oossa · 2026-11-13 · https://oossa.com/en/llama-cpp-update-adds-warnings-for-draft-p-min-usage

The open‑source llama.cpp library released version b12364 on 2026-11-13. It now logs a warning if the --spec-draft-p-min flag is set while using probabilistic draft sampling, suggesting p_min be set to zero. The change skips the p_min cutoff for MTP and DFlash2 drafts, which previously could drop tokens the target model would accept. The warning is only shown for drafters that actually use p_min, avoiding noise for n‑gram and DSpark drafters.

## The facts

- Version b12364 released on 2026-11-13
- Adds warning for --spec-draft-p-min > 0 with probabilistic sampling

## Why it matters

Developers can avoid silent token loss when fine‑tuning draft models, leading to more reliable generation.

## Sources & references

1. [ggml-org/llama.cpp b12364](https://github.com/ggml-org/llama.cpp/releases/tag/b12364) – llama.cpp, 2026-11-13

Last updated: 2026-11-13