GPT OSS 20B — Hardware Requirements & GPU Compatibility
ChatFunctions
GPT-OSS 20B is one of OpenAI's first open-source model releases, marking a historic shift in the company's approach to open weights. At 21.5 billion parameters it delivers strong general-purpose chat and reasoning capabilities informed by the research behind the GPT family, making it a compelling option for users who want OpenAI-grade quality in a locally deployable package. The model runs comfortably on a single high-end consumer GPU such as an RTX 4090 at 4-bit quantization, or on workstation cards with 24 GB or more of VRAM at higher precision. It occupies a practical middle ground between lightweight 7B models and resource-heavy 70B+ offerings.
Runs great on a 24 GB GPU
Runs from 8 GB (IQ2_M · 32K)Best on 24 GB (Q4_K_M · ~190 tok/s · 128K)Macs from 32 GB (~28 tok/s)
Score 96.2#58 of 92 open#15 of 34 that fit 24 GBApache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
llmrun Composite Score →
96.2
- 90% range 93.0–99.3
- #58 of 92 open models
- #15 of 34 that fit 24 GB
Run GPT OSS 20B
Best fit for your Mac Studio M5 Max (36 GB). Every other method, OS and quant is in the step-by-step guide.
Recommended for your Mac Studio M5 Max (36 GB)
mlx_lm.chat --model lmstudio-community/gpt-oss-20b-MLX-8bitMLX repo with its own quantisation
Download: gpt-oss-20b-Q4_K_M.gguf11.62 GBunsloth
Other ways
each opens step-by-stepRun in cloud
Fits on Tesla V100 PCIe 16GB (2 GB headroom) · Q4_K_M
- Generation speed
- ~174 tok/s
- generation speed
- Cost per 1M output tokens
- $0.14
- per 1M output tokens
How Much VRAM Does GPT OSS 20B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On your device | |
|---|---|---|---|---|
| 9.3 GB | 9.8 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 128K context · ~95 tok/s | |
| 9.5 GB | 10 GB | 3-bit small — 3-bit small quantization | Fits: 128K context · ~93 tok/s | |
| 10.6 GB | 11.1 GB | 3-bit medium — 3-bit medium quantization | Fits: 128K context · ~88 tok/s | |
| 10.8 GB | 11.3 GB | 4-bit · legacy — 4-bit legacy quantization | Fits: 128K context · ~87 tok/s | |
| 12.9 GB | 13.4 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 128K context · ~78 tok/s | |
| 15.3 GB | 15.8 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 128K context · ~70 tok/s | |
| 17.6 GB | 18.1 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 128K context · ~63 tok/s | |
| 21.3 GB | 21.8 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 128K context · ~55 tok/s |
Selected Q4_K_M
4-bit medium quantization13.1GBat 16K context
- Weights
- 12.6 GB
- KV cache
- 0.3 GB
- Overhead
- 0.3 GB
Fits Mac Studio M5 Max (36 GB) · 19.4 GB free · ~78 tok/s
≥50 tok/s needs ≥213 GB/s · RTX 5090 ~282, RTX 3090 Ti ~190
Which Hardware Can Run GPT OSS 20B?
Q4_K_M · 12.9 GBGPT OSS 20B (Q4_K_M) requires 12.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 17+ GB is recommended. Using the full 131K context window can add up to 2.2 GB, bringing total usage to 15.1 GB. 26 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 5080. 128 devices with unified memory can run GPT OSS 20B, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, MacBook Pro 14" M3 Pro (18 GB).
16–18 GB
16–18 GB tier →fits · 2.9 GB free · up to 128K ctx
18 GPUs · 1 Mac class · 1 device family
RTX 5080~183Simulate output on RTX 5080RTX 5070 Ti~173Simulate output on RTX 5070 TiRTX 4080 SUPER~149Simulate output on RTX 4080 SUPERM3 Pro*18 GB~34Simulate output on MacBook Pro 14" M3 Pro (18 GB)Orin NX~25Simulate output on NVIDIA Jetson Orin NX 16GBfits · 10.9 GB free · up to 128K ctx
6 GPUs · 2 Mac classes
RTX 3090 Ti~190Simulate output on RTX 3090 TiRTX 4090~190Simulate output on RTX 4090RTX 3090~179Simulate output on RTX 3090M5 Pro24 GB~59Simulate output on Mac Mini M5 Pro (24 GB)M524 GB~35Simulate output on MacBook Air 13" M5 (24 GB)48 GB
48 GB tier →fits · 34.9 GB free · up to 128K ctx
2 Mac classes · 1 device family
M5 Max48 GB~92Simulate output on Mac Studio M5 Max (48 GB)M5 Pro48 GB~59Simulate output on Mac Mini M5 Pro (48 GB)Snapdragon X2 Elite Extreme (X2E-96-100)~50Simulate output on Snapdragon X2 Elite Extreme Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
How good is GPT OSS 20B?
The llmrun Score merges 11 public benchmark boards into a single figure, where 100 represents the average for open models.
Score vs VRAM in the 24 GB tier
- GPT OSS 20B
- Other models
- Best Score for its VRAM
#15 of 34 that fit 24 GBEvery model here fits your Mac Studio M5 Max (36 GB) (36 GB)
Each dot is a model at the best quant that fits a 24 GB GPU. The line joins the models with the best Score for their VRAM; a dot below it is beaten by a model that needs no more memory.
Full 24 GB ranking →Benchmarks
The 6 boards that weigh most in the llmrun Score, of 12 with a result for this model; badges show who ran the test.
- GPQA Diamond
Epoch AI · Aug 2025Independent
Independent60.8%
#41/97 open
- LMArena Text
LMArena · Oct 2026Arena
Arena1317.3
#96/186 open
- AIME 2024/2025
Epoch AI · Aug 2025Independent
Independent65.3%
#33/85 open
- Terminal-Bench
Epoch AI · Aug 2025Lab-reported
Lab-reported3.4%
#16/16 open
- MMLU-Pro
TIGER-LabIndependent
Independent73.6%
#33/144 open
- SciCode
Epoch AI · Aug 2025Lab-reported
Lab-reported34.4%
#47/65 open
In the 24 GB class
34 models fit 24 GBSame Score, only models that fit 24 GB at their best quant: the ranking a 24 GB owner actually chooses from.
- #1
Qwen3.8 27BAlibaba
121.5
~38 tok/s · 17.9 GB
- #2
Qwen3.6 27BAlibaba
111.4
~38 tok/s · 17.9 GB
- #3
Qwen3.6 35B A3BAlibaba
106.9
~151 tok/s · 22.0 GB
- #4
Gemma 4 31B ITGoogle
106.1
~32 tok/s · 21.7 GB
- #5
Qwen3.5 35B A3BAlibaba
105.6
~151 tok/s · 22.0 GB
- #6
Gemma 4 26B A4B ITGoogle
105.4
~171 tok/s · 16.5 GB
Alternatives to GPT OSS 20B
GPT-OSS family · 14 models →Official models with an llmrun Score, each answering one question; numbers are at the best quant that fits a 24 GB GPU.
Runner-up at this size
Phi 4
Microsoft
- 12.4 lower Score
- 0.7 GB less VRAM
- 121 tok/s slower
Score 83.8 · 12.5 GB · ~69 tok/s
Other vendor
Qwen3.8 27B
Alibaba
- 25.3 higher Score
- 4.7 GB more VRAM
- 152 tok/s slower
Score 121.5 · 17.9 GB · ~38 tok/s
Similar models
10| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.6 27BAlibaba | 111.4 | 17.9 GB | ~38 | similar score |
| Qwen3.6 35B A3BAlibaba | 106.9 | 22 GB | ~151 | similar score |
| Gemma 4 31B ITGoogle | 106.1 | 21.7 GB | ~32 | similar score |
| Qwen3.5 35B A3BAlibaba | 105.6 | 22 GB | ~151 | similar score |
| Gemma 4 26B A4B ITGoogle | 105.4 | 16.5 GB | ~171 | similar score |
| Qwen3.5 27BAlibaba | 103.8 | 17.9 GB | ~38 | similar score |
| Seed OSS 36B InstructByteDance-Seed | 102.5 | 22.8 GB | ~31 | similar score |
| Qwen3 30B A3B Thinking 2507Alibaba | 100.6 | 19.4 GB | ~173 | similar score |
| Qwen3.5 9BAlibaba | 99.4 | 6.6 GB | ~103 | similar score |
| Qwen3 32BAlibaba | 98.4 | 22.6 GB | ~32 | similar score |
Specifications
- Publisher
- OpenAI
- Family
- GPT-OSS
- Parameters
- 20.9B
- Active parameters
- 4.2B per token
- Architecture
- GptOssForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 201,088
- Release Date
- 2025-08-04
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does GPT OSS 20B need?
GPT OSS 20B requires 12.9 GB of VRAM at Q4_K_M, or 42.2 GB at BF16. Full 131K context adds up to 2.2 GB (15.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 20.9B × 4.8 bits ÷ 8 = 12.5 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 2.6 GB (at full 131K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 4090 run GPT OSS 20B?
Yes, at Q8_0 (21.3 GB) or lower. Higher quantizations like BF16 (42.2 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
What's the best quantization for GPT OSS 20B?
For GPT OSS 20B, Q4_K_M (12.9 GB) offers the best balance of quality and VRAM usage. Q4_K_L (13.2 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 6.1 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run GPT OSS 20B on a Mac?
Yes, but only at lower quantizations. The smallest Mac that can run GPT OSS 20B is Apple iPhone 17 Pro (12 GB) at IQ2_XXS; 139 of the 156 Macs we list can run it at some quantization. For Q4_K_M (12.9 GB) you need a Mac with more unified memory.
Can I run GPT OSS 20B locally?
Yes — GPT OSS 20B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is GPT OSS 20B?
At Q4_K_M, GPT OSS 20B can reach ~187 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~190 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 2.6 GB active ÷ (8000 × 0.65) = 0.498 ms, plus 24 layers × 0.055 ms = 1.320 ms, so 1000 ÷ 1.818 ms = ~550 tok/s
Estimated speed at Q4_K_M (12.9 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of GPT OSS 20B?
At Q4_K_M, the download is about 12.55 GB. The full-precision BF16 version is 41.83 GB. The smallest option (IQ2_XXS) is 5.75 GB.
Which GPUs can run GPT OSS 20B?
26 consumer GPUs can run GPT OSS 20B at Q4_K_M. Top options include AMD Radeon RX 7900 XT, AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090, AMD Radeon RX 6800. 8 GPUs have plenty of headroom for comfortable inference.
Which devices can run GPT OSS 20B?
128 devices with unified memory can run GPT OSS 20B at Q4_K_M (12.9 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB), Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.