AMD Radeon 780M (Phoenix iGPU)
No editorial image yet — generic vendor mark shown. Credentials in spec table below.
AMD's 780M iGPU (Ryzen 7040/8040 series Phoenix). Shares system RAM via unified memory architecture; 32 GB DDR5 system gives effective 16-20 GB usable for inference. Bandwidth (89 GB/s) is the bottleneck — ~6-12 tok/s on 7B Q4. The 'I have a thin laptop' audience can run AI but slowly.
Sub-scores sum to 181 / 1000. Headline = 181 × 0.70 (Estimated-confidence discount) = 127. This is an algorithmic performance-tier score — distinct from, and often lower than, the editorial “Our verdict” below, which weighs value and real-world fit (especially for hardware we haven’t measured yet). How scoring works →
Extrapolated from 89 GB/s bandwidth — 8.9 tok/s estimated. No measured benchmarks yet.
Plain-English: Doesn't fit modern chat models usefully — vision models won't fit.
Verdicts extrapolated from catalog VRAM + bandwidth + ecosystem flags. Hover any chip for the rationale. Want measured numbers? Submit your own run with runlocalai-bench --submit.
This card is for the operator who already owns a Ryzen 7040/8040 laptop and wants to experiment with local AI without buying a discrete GPU. It runs 7B Q4 models at 6-12 tok/s — usable for chat but too slow for real-time interaction. 13B models dip to ~3-5 tok/s, and anything larger is impractical. The shared memory architecture means the system RAM (ideally 32 GB) is split between GPU and OS, leaving ~16-20 GB for models. What breaks: any model above 13B, any workload requiring sustained throughput, and ROCm support on iGPUs is still rough — expect manual setup and limited compatibility. Pass if you need interactive speeds or plan to run models larger than 7B; a used RTX 3060 12GB will cost ~$200 and deliver 5-10x the performance. Price/value note: the 780M is free with the CPU, so its value is purely incremental — if you already have the laptop, it's a bonus, not a reason to buy.
›Why this rating
The 780M is a capable iGPU for casual local AI experimentation, but its shared memory and low bandwidth severely limit model size and speed. It earns a 3.0 for being a free add-on that can run small models, but it's not a primary inference card.
Overview
AMD's 780M iGPU (Ryzen 7040/8040 series Phoenix). Shares system RAM via unified memory architecture; 32 GB DDR5 system gives effective 16-20 GB usable for inference. Bandwidth (89 GB/s) is the bottleneck — ~6-12 tok/s on 7B Q4. The 'I have a thin laptop' audience can run AI but slowly.
Search-fallback link — editorial hasn't yet curated a retailer URL for this card.
Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.
Specs
| VRAM | 0 GB |
| Power draw (peak) | 28 W |
| Released | 2023 |
| Backends | ROCm Vulkan |
Hardware worth comparing
The closest alternatives by price, memory bandwidth, and form factor, plus a step up and down — so you can frame the buying decision against real options.
Frequently asked
Does AMD Radeon 780M (Phoenix iGPU) support CUDA?
Where next?
Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify hardware specifications.