Model: TeichAI/gpt-oss-20b-claude-4.5-sonnet-high-reasoning-distill-GGUF Source: Original Platform
This model is still in developement.
For the most reliable performance use the following sampling parameters:
temperature
top_k
min_p
top_p
repeat_penalty
Check back later for updates.
This gpt_oss model was trained 2x faster with Unsloth and Huggingface's TRL library.