How fast is gpt-oss-20b on an Mac?

Around 15 tokens per second. Token generation on Apple Silicon is limited by memory bandwidth, and this machine has 68 GB/s. Because this is a mixture-of-experts model only 3.6B of its 20.9B parameters are read per token, which is why it is faster than its size suggests.

How much memory does gpt-oss-20b need?

11.9 GiB for the weights at Q4_K_M, plus 0.4 GiB for an 8K-token KV cache and about 1.4 GiB of runtime overhead — 13.7 GiB in total.

Home › gpt-oss-20b › M1 8GB

Can an M1 8GB Mac run gpt-oss-20b?

M1 · 8 GB unified memory · 68 GB/s · MacBook Air (M1), MacBook Pro 13" (M1)

No — not at Q4_K_M.

gpt-oss-20b needs about 13.7 GiB but an M1 8GB Mac has only ~5.0 GiB available for a model — you are 8.7 GiB short. macOS will swap to SSD and generation will crawl.

Weights

11.9 GiB

20.9B params @ Q4_K_M

KV cache

0.4 GiB

8K context, FP16

Overhead

1.4 GiB

runtime + activations

Total needed

13.7 GiB

of ~5.0 GiB usable

Est. speed

15 tok/s

68 GB/s bandwidth

Max context

n/a

at Q4_K_M

Where the memory goes

Three things occupy memory when a model is loaded: the weights themselves, the KV cache that holds the conversation, and the runtime's own working buffers. Only the first is fixed — the KV cache grows with every token in your context.

Quantisation	Weights	Total	Tok/s	Fits?
FP16	38.9 GiB	42.1 GiB	5	No
Q8_0	20.7 GiB	22.9 GiB	9	No
Q6_K	16.1 GiB	18.0 GiB	11	No
Q5_K_M	13.9 GiB	15.7 GiB	13	No
Q4_K_M	11.9 GiB	13.7 GiB	15	No
Q3_K_M	9.5 GiB	11.1 GiB	19	No
Q2_K	7.3 GiB	8.8 GiB	25	No

The default. Best quality-per-gigabyte for most people.

How long a context fits

The KV cache is often what breaks a setup that looked fine at load time. Quantising it to 8-bit roughly halves its footprint with little measurable quality cost.

Context	KV cache	Total	FP16 KV	Q8 KV
2K	0.1 GiB	13.2 GiB	No	No
4K	0.2 GiB	13.4 GiB	No	No
8K	0.4 GiB	13.7 GiB	No	No
16K	0.8 GiB	14.4 GiB	No	No
32K	1.5 GiB	15.7 GiB	No	No
128K	6.0 GiB	23.8 GiB	No	No

What to run instead on an M1 8GB Mac

These are the largest models that fit comfortably on this machine at Q4_K_M with an 8K context.

Model	Size	Total	Tok/s
Llama 3.2 3B	3.21B	3.6 GiB	28	Yes, comfortably
Qwen2.5 3B	3.09B	2.9 GiB	29	Yes, comfortably
Gemma 2 2B	2.61B	3.0 GiB	34	Yes, comfortably
Qwen2.5 1.5B	1.54B	1.9 GiB	58	Yes, comfortably
Llama 3.2 1B	1.24B	1.8 GiB	72	Yes, comfortably

Macs that run gpt-oss-20b comfortably

The smallest configuration that handles it is the M2 24GB (MacBook Air 15" (M2)).

Mac	Memory	Bandwidth	Tok/s
M2 24GB	24 GB	100 GB/s	23
M4 24GB	24 GB	120 GB/s	27
M4 Pro 24GB	24 GB	273 GB/s	62
M1 Max 32GB	32 GB	400 GB/s	91
M2 Max 32GB	32 GB	400 GB/s	91
M4 32GB	32 GB	120 GB/s	27
M3 Max 36GB	36 GB	300 GB/s	68
M4 Max 36GB	36 GB	410 GB/s	93
M4 Pro 48GB	48 GB	273 GB/s	62
M4 Max 48GB	48 GB	546 GB/s	124
M1 Max 64GB	64 GB	400 GB/s	91
M2 Ultra 64GB	64 GB	800 GB/s	181

Common questions

Can an M1 8GB Mac run gpt-oss-20b?: No. At Q4_K_M it needs about 13.7 GiB, and an M1 8GB Mac has roughly 5.0 GiB available for a model after macOS takes its share.
How fast is gpt-oss-20b on an Mac?: Around 15 tokens per second. Token generation on Apple Silicon is limited by memory bandwidth, and this machine has 68 GB/s. Because this is a mixture-of-experts model only 3.6B of its 20.9B parameters are read per token, which is why it is faster than its size suggests.
How much memory does gpt-oss-20b need?: 11.9 GiB for the weights at Q4_K_M, plus 0.4 GiB for an 8K-token KV cache and about 1.4 GiB of runtime overhead — 13.7 GiB in total.