Search: “ai2”

Grid Timeline Evolution

3 results3 results

Model Releases

Four small architecture decisions can cost up to 47% of a model's long-context performance. New research from Ai2, Carnegie Mellon, and the …

DGX agent

Four small architecture decisions can cost up to 47% of a model's long-context performance. New research from Ai2, Carnegie Mellon, and the University of Washington isolates them. Normalization, GQA,

model-releasesdair-ai--x

12 Aug 2026

DGX agent

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Ai2 just released TMax 27B on Hugging Face A 27B terminal agent that hits 42.7% on Terminal Bench 2.0, rivaling models 40× its size.

DGX agent

AI2 released TMax 27B, a 27 billion parameter terminal agent model available on Hugging Face that achieves 42.7% performance on Terminal Bench 2.0, matching the capabilities of much larger models desp

model-releasesclem-delangue--x

22 Jun 2026

Agents

seattle friends! 📅 i'm doing a talk on @activegraphai on june 24th at the AI house (@ai2incubator), come by :) https://luma.com/aihouse-2fd…

DGX agent

Yohei Nakajima announced a talk about ActiveGraph AI scheduled for June 24th at the AI House, an AI2 incubator venue in Seattle, with an invitation for local attendees to participate. The event detail

agentsyohei-nakajima--x

8 Jun 2026

3 results