GPT OSS 20B Heretic Ara v3 — Hardware Requirements & GPU Compatibility
Chat
GPT OSS 20B Heretic Ara v3 is a 1.8B-parameter open language model from p-e-w in the GPT-OSS family. It supports a context window of up to 131,072 tokens. At IQ4_NL it needs about 1.39 GB of VRAM — see which GPUs and Macs can run it below.
1.1K downloads
21 likesSpecifications
- Publisher
- p-e-w
- Family
- GPT-OSS
- Parameters
- 1.8B
- Architecture
- GptOssForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 201,088
- Release Date
- 2026-03-05
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does GPT OSS 20B Heretic Ara v3 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| IQ4_NL | 4.50 | 1.4 GB | 5.8 GB | 1.02 GB | Importance-weighted 4-bit, non-linear |
| Q5_1 | 5.50 | 1.6 GB | 6.1 GB | 1.24 GB | 5-bit legacy quantization with offset |
| Q8_0 | 8.00 | 2.2 GB | 6.6 GB | 1.80 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run GPT OSS 20B Heretic Ara v3?
Q8_0 · 2.2 GBGPT OSS 20B Heretic Ara v3 (Q8_0) requires 2.2 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 3+ GB is recommended. Using the full 131K context window can add up to 4.4 GB, bringing total usage to 6.6 GB. 35 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run GPT OSS 20B Heretic Ara v3?
Q8_0 · 2.2 GB33 devices with unified memory can run GPT OSS 20B Heretic Ara v3, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomRelated Models
Frequently Asked Questions
- How much VRAM does GPT OSS 20B Heretic Ara v3 need?
GPT OSS 20B Heretic Ara v3 requires 1.4 GB of VRAM at IQ4_NL, or 2.2 GB at Q8_0. Full 131K context adds up to 4.5 GB (5.8 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 1.8B × 4.5 bits ÷ 8 = 1 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 4.8 GB (at full 131K context)
VRAM usage by quantization
IQ4_NL1.4 GBIQ4_NL + full context5.8 GB- What's the best quantization for GPT OSS 20B Heretic Ara v3?
For GPT OSS 20B Heretic Ara v3, Q5_1 (1.6 GB) offers the best balance of quality and VRAM usage. Q8_0 (2.2 GB) provides better quality if you have the VRAM. The smallest option is IQ4_NL at 1.4 GB.
VRAM requirement by quantization
IQ4_NL1.4 GBQ5_1 ★1.6 GBQ8_02.2 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run GPT OSS 20B Heretic Ara v3 on a Mac?
GPT OSS 20B Heretic Ara v3 requires at least 1.4 GB at IQ4_NL, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run GPT OSS 20B Heretic Ara v3 locally?
Yes — GPT OSS 20B Heretic Ara v3 can run locally on consumer hardware. At IQ4_NL quantization it needs 1.4 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is GPT OSS 20B Heretic Ara v3?
At IQ4_NL, GPT OSS 20B Heretic Ara v3 can reach ~2097 tok/s on AMD Instinct MI300X. On NVIDIA GeForce RTX 4090: ~471 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 1.4 × 0.55 = ~2097 tok/s
Estimated speed at IQ4_NL (1.4 GB)
~2097 tok/s~471 tok/s~1568 tok/s~1297 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of GPT OSS 20B Heretic Ara v3?
At IQ4_NL, the download is about 1.02 GB. The full-precision Q8_0 version is 1.80 GB.
- Which GPUs can run GPT OSS 20B Heretic Ara v3?
35 consumer GPUs can run GPT OSS 20B Heretic Ara v3 at IQ4_NL (1.4 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 35 GPUs have plenty of headroom for comfortable inference.
- Which devices can run GPT OSS 20B Heretic Ara v3?
33 devices with unified memory can run GPT OSS 20B Heretic Ara v3 at IQ4_NL (1.4 GB), including Mac Mini M4 (16 GB), Mac Mini M4 (32 GB), Mac Mini M4 Pro (24 GB), Mac Mini M4 Pro (48 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.