MemT 1.0 open-sourced with 36B and 288B models, reaching near-300B memory-native validation

MemT 1.0 open-sourced with 36B and 288B models, reaching near-300B memory-native validation

N
News Editor
2026-10-20 22:55:54
MemT 1.0, a memory-native model series, has been released as open source with two versions: the 36B-parameter Nano model and the 288B-parameter Flash model. According to the release cited by Techub News, the series adds a parallel memory-compute path inside Transformer modules, allowing experience to be encoded as internal dynamic states and read or written through training. The team said this marks the first engineering validation of a memory-native design at nearly 300B parameters. On benchmarks including SWE-bench Verified, MemT-1.0-Nano posted a 69.73% average success rate on key breakpoints. The share of tasks that succeeded in all three runs rose from 54.00% under a no-experience baseline to 62.80%. In tasks such as maze exploration, the model was able to use experience from the previous round to improve later actions. MemT 1.0 uses Gated DeltaNet-2 for per-token memory updates and introduces a Mixture-of-Storage approach built on a Mixture of Experts, or MoE, architecture to expand memory capacity. The team trained the memory system on 34 billion tokens across continued pretraining and post-training stages.

MemT 1.0, a memory-native model series, has been released as open source, according to Techub News. The release includes the 36B-parameter Nano model and the 288B-parameter Flash model.

The series adds a parallel memory-compute path to Transformer modules. It encodes experience into internal dynamic states and learns to read and write that memory through training. The team said this is the first engineering validation of a memory-native system at nearly 300B parameters.

Benchmark results

On benchmarks including SWE-bench Verified, MemT-1.0-Nano recorded a 69.73% average success rate on key breakpoints. The share of tasks that succeeded in all three runs increased from 54.00% under a no-experience baseline to 62.80%.

In tasks such as maze exploration, the model was able to use experience from the previous round to improve later actions.

Architecture and training

MemT 1.0 uses Gated DeltaNet-2 for per-token memory updates. It also introduces a Mixture-of-Storage design based on a Mixture of Experts, or MoE, architecture to expand memory capacity.

The team trained the memory system on 34 billion tokens, covering both continued pretraining and post-training stages.

The item cited Quantum Bit.

This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan.
7700

Disclaimer:

The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.

Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.

MemT 1.0 open-sourced with 36B and 288B models, reaching near-300B memory-native validation | Bit.Fan