Daily AI news, sourcedOn Oct 20, MemTensor released MemT-1.0, open-weight models with memory built into the model.
Where it stands
On Oct 20, MemTensor released MemT-1.0, open-weight models with memory built into the model.
Timeline · newest first
- Official
MemTensor releases MemT-1.0, open-weight models with memory built into the model
MemTensor published this on Oct 20, 2026 (source: huggingface.co).
MemTensor released MemT-1.0, open-weight models that build persistent memory into the model's computation so past information can be reused without reloading the original context. The Flash version is based on DeepSeek-V4-Flash under an MIT license, and Nano is based on Qwen3.6-35B-A3B. MemTensor says it leads parametric memory methods on 11 memory benchmarks; these are its own results.
Source: MemTensor
Related stories
NVIDIA53 updates · Models→Claude37 updates · Models→Google Gemini37 updates · Models→Qwen23 updates · Models→
How we label
- Official
- An institution announces something that has already happened, through its own channel or a founder or executive account.
- Statement by …
- A named person's views, predictions, plans or comments, including remarks reported by media when the original is unavailable.
- Reported by …
- Only media or another third party is saying it. We name the outlet.
- Unconfirmed
- Circulating online (posts, comments, leaks). No company, named person or outlet has confirmed it.
Items older than 72 hours never go in Top stories and show their original date.