What AI models can NVIDIA RTX 4080 Laptop (120-150W) run?

The NVIDIA RTX 4080 Laptop (120-150W) can run 40 AI models. Top performers include all-MiniLM-L6-v2, nomic-embed-text v1.5, Llama 3.2 1B Instruct. See the full compatibility table above for speeds and quality ratings.

Is NVIDIA RTX 4080 Laptop (120-150W) good for AI coding?

With 12 GB, the NVIDIA RTX 4080 Laptop (120-150W) runs 7-8B coding models at the Starter tier. Good for basic code completion.

How much VRAM does NVIDIA RTX 4080 Laptop (120-150W) have?

The NVIDIA RTX 4080 Laptop (120-150W) has 12 GB of GDDR6 VRAM with 384 GB/s bandwidth.

Can NVIDIA RTX 4080 Laptop (120-150W) run 70B models?

70B models can run on the NVIDIA RTX 4080 Laptop (120-150W) with CPU offloading, but performance will be reduced. Consider a device with 48GB+ inference memory for full-speed 70B inference.

Is NVIDIA RTX 4080 Laptop (120-150W) worth it for AI?

Pricing for NVIDIA RTX 4080 Laptop (120-150W) has not been announced. It offers 12 GB GDDR6 VRAM, but OwnRig should treat recommendations as provisional until pricing and benchmarks are available.

Laptop GPU

NVIDIA RTX 4080 Laptop (120-150W)

12 GB GDDR6 · 384 GB/s

Pricing

Included in laptop

Not sold as a standalone component

VRAM

12 GB

Bandwidth

384 GB/s

TDP

120W

Models

Tier

Starter

The NVIDIA RTX 4080 Laptop (120-150W) with 12 GB GDDR6 VRAM can handle 40 AI models across embedding, ai_building, coding. Best performance: all-MiniLM-L6-v2 at 8400 tok/s (excellent). For AI coding workflows, it supports the Starter AI Coding tier, good for 7–8B models. Current price has not been announced.

Source: OwnRig methodology

VRAM

12 GB

Bandwidth

384 GB/s

Memory Type

GDDR6

TDP

120W

Form Factor

Laptop (soldered)

Laptop Performance Note

Laptop GPU performance varies by manufacturer, cooling design, and power limits. The tok/s numbers below reflect sustained performance after thermal throttling, not peak. Actual results on your specific laptop may differ by 10-20%.

Builder Capability: Starter AI Coding

Runs 7-8B models comfortably. Good for basic local code completion and small model experiments.

Software

Inference Backends

The software stacks that matter most for real-world inference on this device.

CUDA

production

Primary high-performance backend for NVIDIA inference workloads.

Vulkan

stable

Fallback backend for llama.cpp and related local runtimes.

What it can run

40 models


all-MiniLM-L6-v2	FP16	8400 tok/s	Excellent
Arcee Trinity Mini 26B	Q3_K_M	6 tok/s	Not viable
Arcee Trinity Nano 6B	Q8_0	68 tok/s	Excellent
Codestral 22B	Q3_K_M	8 tok/s	Marginal
DeepSeek Coder V2 Lite 16B	Q4_K_M	39 tok/s	Good
DeepSeek R1 Distill Qwen 7B	Q5_K_M	34 tok/s	Good
Gemma 2 9B Instruct	Q5_K_M	34 tok/s	Good
Gemma 3 12B	Q4_K_M	22 tok/s	Acceptable
Gemma 3 4B	Q8_0	59 tok/s	Excellent
Gemma 4 26B-A4B	Q3_K_M	8 tok/s	Not viable
Gemma 4 E2B	Q8_0	57 tok/s	Good
Gemma 4 E4B	Q8_0	35 tok/s	Good
GigaChat Lightning 10B	Q4_K_M	80 tok/s	Acceptable
InternLM 2.5 7B Chat	Q5_K_M	32 tok/s	Good
Llama 3.1 8B Instruct	Q5_K_M	36 tok/s	Good
Llama 3.2 1B Instruct	Q8_0	98 tok/s	Excellent
Llama 3.2 3B Instruct	Q8_0	67 tok/s	Excellent
LLaVA 1.6 13B	Q4_K_M	20 tok/s	Acceptable
Mistral 7B Instruct v0.3	Q5_K_M	35 tok/s	Good
nomic-embed-text v1.5	Q8_0	4550 tok/s	Excellent
NVIDIA Nemotron-3-super-120B-A12B	Q2_K	–	Not viable
Phi-3 Medium 14B Instruct	Q3_K_M	25 tok/s	Good
Phi-3 Mini 3.8B Instruct	Q8_0	55 tok/s	Excellent
Phi-4 14B	Q3_K_M	24 tok/s	Acceptable
Phi-4 Mini	Q8_0	57 tok/s	Excellent
Qwen 2.5 14B Instruct	Q4_K_M	21 tok/s	Acceptable
Qwen 2.5 7B Instruct	Q5_K_M	34 tok/s	Good
Qwen 2.5 Coder 7B Instruct	Q5_K_M	34 tok/s	Good
Qwen3-14B Instruct	Q5_K_M	16 tok/s	Acceptable
Qwen3-8B Instruct	Q8_0	21 tok/s	Acceptable
Qwen3.5-122B-A10B	Q3_K_M	–	Not viable
Qwen3.5-27B	Q3_K_M	8 tok/s	Marginal
Qwen3.5-397B (MoE)	Q2_K	–	Not viable
Qwen3.6-27B	Q3_K_M	–	Not viable
Stable Diffusion 3 Medium	FP16	–	Excellent
Stable Diffusion 3.5 Large	Q8_0	–	Good
Stable Diffusion XL 1.0	FP16	–	Excellent
StarCoder 2 15B	Q3_K_M	20 tok/s	Acceptable
Whisper Large V3	FP16	–	Excellent
Whisper Large V3 Turbo	FP16	–	Excellent

Showing 40 of 40 entries

Looking for a desktop build?

Desktop GPUs offer higher sustained performance with no thermal throttling. Check our curated desktop builds for dedicated AI workstations.

FAQ

Frequently Asked Questions

What AI models can NVIDIA RTX 4080 Laptop (120-150W) run?: The NVIDIA RTX 4080 Laptop (120-150W) can run 40 AI models. Top performers include all-MiniLM-L6-v2, nomic-embed-text v1.5, Llama 3.2 1B Instruct. See the full compatibility table above for speeds and quality ratings.
Is NVIDIA RTX 4080 Laptop (120-150W) good for AI coding?: With 12 GB, the NVIDIA RTX 4080 Laptop (120-150W) runs 7-8B coding models at the Starter tier. Good for basic code completion.
How much VRAM does NVIDIA RTX 4080 Laptop (120-150W) have?: The NVIDIA RTX 4080 Laptop (120-150W) has 12 GB of GDDR6 VRAM with 384 GB/s bandwidth.
Can NVIDIA RTX 4080 Laptop (120-150W) run 70B models?: 70B models can run on the NVIDIA RTX 4080 Laptop (120-150W) with CPU offloading, but performance will be reduced. Consider a device with 48GB+ inference memory for full-speed 70B inference.
Is NVIDIA RTX 4080 Laptop (120-150W) worth it for AI?: Pricing for NVIDIA RTX 4080 Laptop (120-150W) has not been announced. It offers 12 GB GDDR6 VRAM, but OwnRig should treat recommendations as provisional until pricing and benchmarks are available.

Own this GPU?

See every AI model it supports, expected performance, and how to build around it.

Check my rig

All GPUs