What are the top AI Models that are still relevant today from 2024 and 2025?
In 2026, the AI landscape has seen a staggering evolution, with Mixture of Experts (MoE) models becoming prevalent. Dense models are largely obsolete, though some exceptions remain. The discussion highlights the ability to handle large parameter models, indicating a preference for powerful AI systems. The Kimi k3 model is specifically mentioned as being out of contention for large quantization.
This discussion, unlike typical forward-looking analyses, reflects on the relevance of 2024 and 2025 AI models from the perspective of 2026, highlighting the shift to MoE models.
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
PublishedOffset at this time: UTC+0Sep 13, 2026, 12:23 UTC
IngestedOffset at this time: UTC+0Sep 13, 2026, 15:01 UTC
- Published
- Sep 13, 2026, 12:23
- Ingested
- Sep 13, 2026, 15:01
- Source type
- Dev community
- Tier
- Community
- Source status
- Healthy
Tier is a per-source editorial setting, not a per-item score.
We are in 2026 and it has been a crazy year. The evolution of the technology has been staggering to say the least. Pretty much we are on the MoE period and dense models are almost in the way of the dodo except for a few.
I was just thinking, are there any models of the last 2 years that you would fire up and still feel useful? Example Deep Seek R1.
If you have recommendations write them down. I have 28TB of storage and I am backing up relevant models that can still be useful, from current to older ones. But clearly not all are worth keeping.
Noted: I can hold large parameter models therefore not limited to small ones. No Kimi k3 at large quant thats out of the question lol