Huihui GPT OSS 20B BF16 Abliterated vs GPT OSS 20B RichardErkhov Heresy

Side-by-side comparison of VRAM requirements, quantization, context length, and hardware compatibility.

Specifications

Huihui GPT OSS 20B BF16 AbliteratedGPT OSS 20B RichardErkhov Heresy
Parameters20.9B21.5B
Context131K131K
ArchitectureGptOssForCausalLMGptOssForCausalLM
LicenseApache 2.0Apache 2.0
Downloads30.1K181
ReleasedSep 2025Feb 2026

VRAM by Quantization: Huihui GPT OSS 20B BF16 Abliterated vs GPT OSS 20B RichardErkhov Heresy

QuantizationBitsHuihui GPT OSS 20B BF16 Abliterated VRAMGPT OSS 20B RichardErkhov Heresy VRAM
Q2_K3.409.5 GB
Q3_K_M3.9010.9 GB
Q3_K_S3.509.8 GB
Q4_04.0011.1 GB
Q4_K_M4.8013.3 GB
Q5_K_M5.7015.7 GB
Q6_K6.6018.1 GB
Q8_08.0021.3 GB21.9 GB

Verdict

Huihui GPT OSS 20B BF16 Abliterated needs less VRAM at IQ4_NL (12.1 GB vs 12.5 GB), so it fits on smaller GPUs. Huihui GPT OSS 20B BF16 Abliterated is the more widely downloaded of the two.

Frequently Asked Questions

Which needs less VRAM, Huihui GPT OSS 20B BF16 Abliterated or GPT OSS 20B RichardErkhov Heresy?

At IQ4_NL, Huihui GPT OSS 20B BF16 Abliterated needs 12.1 GB and GPT OSS 20B RichardErkhov Heresy needs 12.5 GB, so Huihui GPT OSS 20B BF16 Abliterated is the lighter option to run locally.

Which has a longer context window, Huihui GPT OSS 20B BF16 Abliterated or GPT OSS 20B RichardErkhov Heresy?

Huihui GPT OSS 20B BF16 Abliterated supports 131,072 tokens and GPT OSS 20B RichardErkhov Heresy supports 131,072 tokens.

What is the difference between Huihui GPT OSS 20B BF16 Abliterated and GPT OSS 20B RichardErkhov Heresy?

Huihui GPT OSS 20B BF16 Abliterated is a 20.9B model from huihui-ai (GPT-OSS family), while GPT OSS 20B RichardErkhov Heresy is a 21.5B model from MuXodious (GPT-OSS family). Compare their VRAM requirements above to see which fits your GPU or Mac.