gpt-oss-20b
Q4 GGUF download: unsloth/gpt-oss-20b-GGUF - 11.62 GB on disk, read 2026-09-27. VRAM figures above are estimates; this is the actual file size.
Hosted API cost & speed
What it costs to rent this model instead of running it locally
| Metric | Value |
|---|---|
| Input (per 1M tokens) | $0.07 |
| Output (per 1M tokens) | $0.30 |
| Median speed (hosted API) | 259 tok/s |
Median across hosted providers, measured by Artificial Analysis - updated 2026-11-16. Local llama.cpp speed depends on your hardware — see the hardware pages for estimates.
Arena rating (LMArena)
Human preference wins from head-to-head chat battles
| Metric | Value |
|---|---|
| Overall arena rating | 1287.6 |
| Rank (overall) | #264 |
| Battle votes | 10,400 |
Overall rating published by LMArena on 2026-11-09 - a pairwise human-preference Elo across live chat battles; higher is better. Fetched 2026-11-12. Open-weights models compete head-to-head with proprietary ones here — no benchmark prompts, just what people pick.
Standard benchmarks
Performance across standard evaluations
| Benchmark | Score |
|---|---|
| MMLU-Pro | 74.8 |
| GPQA Diamond | 61.1 |
| AIME | 62.3 |
| SWEBENCH | 37.4 |
| HLE | 5.3 |
| LiveCodeBench | 65.2 |
| IFBench | 57.8 |
| Terminal-Bench | 4.5 |
Source: Artificial Analysis component evals - updated 2026-11-16. Additional scores from Hugging Face OpenEvals official leaderboards - updated 2026-11-16.
Will it run on your hardware?
Pick your GPU memory - see which quantizations fit, and the cheapest card for the rest
Need an exact figure for your context length? Use the VRAM calculator.
Run it locally
Copy-paste - running in under a minute
vllm serve openai/gpt-oss-20bNew to this? Start with Ollama · serve to many users with vLLM.
Deep dive
Notes, sources, and the full write-up
gpt-oss-20bLink to this section
gpt-oss-20b is a 21.5B-parameter apache-2.0 model from openai. At Q4_K_M it needs roughly 12 GB of VRAM, placing it in the 8–12 GB GPU hardware tier.
SpecificationsLink to this section
| Spec | Value |
|---|---|
| Parameters | 21.5B |
| Context length | 131K tokens |
| License | apache-2.0 |
| Modalities | text |
| Released | 2025-08-04 |
| Weights | openai/gpt-oss-20b |
VRAM requirementsLink to this section
| Quant | VRAM | Runs on |
|---|---|---|
| Q4_K_M | ~12 GB | RTX 3060 12GB, RTX 4070 |
| Q5_K_M | ~15 GB | RTX 4060 Ti 16GB, RTX 4080 |
| Q8_0 | ~23 GB | RTX 3090, RTX 4090 |
| FP16 | ~43 GB | RTX 6000 Ada, dual RTX 3090 |
VRAM is estimated from parameter count; MoE models still need all weights resident.
How to runLink to this section
vLLM:
vllm serve openai/gpt-oss-20bPopularityLink to this section
gpt-oss-20b has 7,780,249 downloads in the last month on HuggingFace and 4,682 likes.
From the community
What users are saying
“I have the suspicion that this is how GPT-OSS-20B would generate a visualization of it's embeddings. Happy to learn otherwise.”
“GPT-OSS-20B-Vision: First community VLM for GPT-OSS, trained on a single DGX Spark A couple weeks ago I shipped an MCP server (noapi-google-search-mcp) and people in the community challenged me to do something harder - build a VLM.”
“They are incomparable. Token embeddings generated with something like word2vec worked well because the networks are shallow and therefore the learned semantic data can be contained solely and independently within the embeddings themselves. Token embeddings as a part of an LLM (e.g.”
Frequently asked
Quick answers to common questions
How much VRAM does gpt-oss-20b need?
gpt-oss-20b with 21.5B parameters needs approximately 12 GB at Q4_K_M quantization. Use our VRAM calculator for an exact estimate.
Is gpt-oss-20b better than other openai models?
gpt-oss-20b has 21.5B parameters with 131,072 context - a strong choice for long-context, general-chat.
What license is gpt-oss-20b under?
gpt-oss-20b is released under the apache-2.0 license, making it suitable for most commercial and personal projects.
What hardware runs gpt-oss-20b well?
With 21.5B parameters, gpt-oss-20b requires adequate VRAM. High-end GPUs like the RTX 4090 (24GB), RTX 5090 (32GB), or Mac Studio with unified memory are good options. Check our hardware directory for specific recommendations.
What is the best quantization for gpt-oss-20b?
Q4_K_M is the recommended sweet spot - ~98% of FP16 quality at ~27% of the size. Q5_K_M (~15 GB) is an option if you have spare VRAM. Use our VRAM calculator to compare.
How long can gpt-oss-20b's context window handle?
gpt-oss-20b supports a 131,072-token context window - enough for very long documents, codebases, or multi-turn conversations. Real-world usable context may vary by implementation.
What models compete with gpt-oss-20b?
gpt-oss-20b competes with other models in its class. Browse our model directory for comparisons, benchmarks, and community reviews to find the best fit.
Can my GPU run gpt-oss-20b?
Fit and estimated speed on popular hardware
Nearby options
Similar models and compatible hardware by spec