What is the biggest model the RTX 3090 can run?

Qwen3 32B, which needs about 22 GB, on the 24 GB version at about 26 tok/s.

What is the biggest model the Tesla P40 can run?

Qwen3 32B, which needs about 22 GB, on the 24 GB version at about 10 tok/s.

RTX 3090 vs Tesla P40 for local AI

Q: Which is better for local AI, the RTX 3090 or the Tesla P40?

The RTX 3090, for speed. They fit the same models. Both fit the same models we track, up to Qwen3 32B. The RTX 3090 generates about 2.7× as fast on any model both can run, because it has 936 GB/s of memory bandwidth to the Tesla P40's 346 GB/s.

Both fit the same models we track, up to Qwen3 32B. The RTX 3090 generates about 2.7× as fast on any model both can run, because it has 936 GB/s of memory bandwidth to the Tesla P40's 346 GB/s.

The numbers that matter

Memory and bandwidth

Memory decides which models fit. Bandwidth decides how fast they generate, because every token reads the model's active weights once.

	RTX 3090	Tesla P40
Memory	24 GB VRAM	24 GB VRAM
Bandwidth	936 GB/s	346 GB/s
Biggest model that fits	Qwen3 32B	Qwen3 32B
Best everyday pick	Qwen3 32B, ~26 tok/s	Qwen3 30B-A3B, ~61 tok/s

Model by model

Every model, side by side

Estimated generation speed at the largest memory size each was sold in (24 GB and 24 GB).

Model	RTX 3090	Tesla P40
Qwen3 4B About 4 GB	~210 tok/s Runs great	~76 tok/s Runs great
Qwen3 8B About 7 GB	~99 tok/s Runs great	~37 tok/s Runs great
Gemma 3 12B About 10 GB	~64 tok/s Runs great	~23 tok/s Usable
Qwen3 14B About 11 GB	~55 tok/s Runs great	~20 tok/s Usable
gpt-oss 20B About 16 GB	~160 tok/s Runs great	~61 tok/s Runs great
Gemma 3 27B About 19 GB	~30 tok/s Runs great	~11 tok/s Usable
Qwen3 30B-A3B About 21 GB	~160 tok/s Runs great	~61 tok/s Runs great
Qwen3 32B About 22 GB	~26 tok/s Usable	~10 tok/s Slow
Llama 3.3 70B About 45 GB	Needs 45 GB	Needs 45 GB
gpt-oss 120B About 67 GB	Needs 67 GB	Needs 67 GB
Qwen3 235B-A22B About 144 GB	Needs 144 GB	Needs 144 GB
DeepSeek R1 671B About 406 GB	Needs 406 GB	Needs 406 GB

Which should you get?

The RTX 3090, for speed. They fit the same models. Price matters as much as either number: check used prices on the builds page, where every parts list has a measured or estimated speed.

How we estimate

Both columns use the same formula: memory bandwidth divided by the weights read per token, calibrated on our own rack, with Ollama's default 4-bit sizes. Real results are usually within about a third. Measured either one? Send us the number.

Compare

Other matchups

Every comparison