The debate about 8GB VRAM GPUs rolls on against a backdrop of overpriced memory chips. And while 8GB is still enough for the vast majority of games at 1080p, its days are surely dwindling. Yet, because Nvidia shows no sign of announcing a new family of graphics cards (“Rubin who?”), we asked ChatGPT. What does its ideal RTX 60-Series graphics card lineup look like?

ChatGPT’s ideal ‘RTX 60-series’ lineup

1. A single 8GB VRAM card

There’s a lot we could analyze, but this is a speculative list with some clear (theoretical) highlights. So let’s start with the headline. The proposed RTX 6050 is the only 8GB VRAM graphics card here. An RTX 6050, costing $249 MSRP but capable of entry-level 1080p gaming, isn’t terrible. That’s assuming that there’s a decent jump in generational performance.

No, it’s not particularly inspiring, and 8GB cards do live on in ChatGPT’s family. But only briefly. And the alternative is pulling up the entrance ramp for desktop gaming. I’d much rather a family with a sub-$300 option than not. A card for the first-gaming-PC users, the Esports players… let’s not completely scorch the earth of the closest thing we have to affordable options.

2. Mid-range clarity

Further up in the lineup, after a clear jump in performance to an RTX 6060 with 12GB, ChatGPT makes a great decision in model clarity. A singular RTX 6060 Ti model with 16GB of VRAM. No 8GB variant in sight, and a clear purpose: 1440p High gaming for $449. This model allows for a $30 jump in MSRP compared to the RTX 5060 Ti 16GB. Is it particularly affordable? No, but when your base 6060 model packs 12GB VRAM, that makes more sense for a variant with more memory.

Then comes what would surely be a big win in performance and flexibility: a 16GB mid-range RTX 6070. A sight to behold: a clear step up from the RTX 5070 and 12GB, although with an expected price hike. No more mid-range 12GB cards, please. Let the 1440p gamers enjoy more headroom and a jump from a 192-bit to a 256-bit memory bus. Filling out the mid-range is a 4K entry with an RTX 6070 Ti 20GB. Another sturdy step up.

3. 4K gaming versus AI tasks

At the top end of the ChatGPT-proposed RTX 60-Series, there’s clear separation and further logic. We’ve ranked the RTX 5080 as the best graphics card for gaming because it handles 4K gaming easily, while the RTX 5090 isn’t necessary for most. ChatGPT’s RTX 6080 and 6080 Ti would cover most of the high-end consumer market, while an RTX 6090 and RTX 6090 Ti (Titan Class) would support more advanced AI and prosumer workloads.

Once more, the VRAM here packs a lot of theoretical punch. Compared to the RTX 5080 and its 16GB of DDR7 VRAM, a massive 24GB and (up to) a 384-bit memory bus (6080 Ti) read like near-flagship models already. Add on the 6090 and the 6090 Ti – “A niche halo product for creators, researchers, and enthusiasts” – and the lineup is complete.

ChatGPT’s goals for this lineup

Handily, ChatGPT explained its homework and made lots of compelling points, stating: “Instead of the uneven jumps that sometimes happen between generations, each step would ideally deliver around a 20–30% improvement over the model below it.” Beyond this, the AI tool suggested that “an ideal RTX 6000 series would also address some common requests”. Such as:

• More VRAM across the board (no 8GB cards above the entry level).
• Wider memory buses on the mid-range to avoid bandwidth bottlenecks.
• Better performance per watt than the previous generation.
• Lower idle power consumption, especially for multi-monitor setups.
• Stronger ray tracing hardware with less reliance on upscaling.
• A new generation of AI upscaling and frame generation that produces fewer artifacts and lower latency.
• DisplayPort 2.1 UHBR20 support across the entire lineup.
• Improved hardware encoding and decoding for AV1 and future codecs.

The one thing missing here? Clear improvements on pricing. But with memory costs likely to remain high for some time, it’s a lot to hope for. Still, at least GPU VRAM capacity was addressed. And ChatGPT had a future-facing sign-off: “As games, content creation, and AI workloads continue to grow, this configuration would give each card a longer useful lifespan and reduce the chance that memory capacity – not GPU compute power – becomes the limiting factor.” Yes, quite a shared sentiment.

About the Author

Kevin Pocock

Kevin has a broad interest and enthusiasm for consumer electronics, PCs, and all things consumer tech. And more than 17 years experience in tech journalism.