Taira Anderson
Taira Anderson is a project manager at the Allen Institute for AI (Ai2), where they work on the lab's fully open Olmo language models and Molmo vision-language models. They coauthored Molmo and PixMo and OLMo 2 in 2024, then OLMoTrace and Olmo 3 in 2025. In 2026, they managed the Molmo2 project and coauthored the MolmoWeb web agents and MolmoMotion, a 3D motion-forecasting model. As of October 2026, they are a Project Manager at Ai2 on the NSF OMAI project, a national-level open AI infrastructure effort built on Olmo and Molmo.
Timeline
-
September 2024
Molmo and PixMo
In September 2024, they appear among the authors of "Molmo and PixMo," which presents Molmo, a family of open-weight vision-language models, and PixMo, a collection of new datasets for pre-training and fine-tuning.
-
December 2024
OLMo 2 open language models
In December 2024, they coauthor "2 OLMo 2 Furious," which presents OLMo 2, a family of fully open language models at 7B, 13B and 32B scales. The release includes model weights, full training data, training code and recipes.
-
April 2025
The OLMoTrace system
In April 2025, they coauthor OLMoTrace, a system that traces the outputs of language models back to their multi-trillion-token training data in real time. The system is publicly available and fully open-source.
-
December 2025
The Olmo 3 model family
Olmo 3 arrives in December 2025 as a family of fully open language models at the 7B and 32B parameter scales, and they are among its authors. The release covers every stage, checkpoint, data point and dependency used to build the models.
-
January 2026
Managing the Molmo2 project
Molmo2, a family of open-weight vision-language models for video understanding and point-driven grounding, appears in January 2026. They are a coauthor and manage the project.
-
April 2026
MolmoWeb open web agents
In April 2026, they coauthor MolmoWeb, a family of fully open multimodal web agents in 4B and 8B sizes, together with MolmoWebMix, a mixture of browser task demonstrations and web-GUI perception data.
-
June 2026
MolmoMotion 3D motion forecasting
MolmoMotion, released in June 2026, predicts the future 3D trajectory of each query point on an object from a short visual history and a language description of the goal. They are a coauthor, with the Allen Institute for AI as their affiliation.
-
As of October 2026
Project Manager on NSF OMAI
They are a Project Manager at Ai2 on the NSF OMAI team, a project supported by the U.S. National Science Foundation and NVIDIA to create a national-level AI infrastructure built by Ai2. NSF OMAI builds on Ai2's fully open models Olmo and Molmo.
Sources and credits
Photo: We have no published photo of Taira Anderson yet. Send us a free-licensed photo.
- Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models
- GitHub - allenai/molmo: Code for the Molmo Vision-Language Model
- 2 OLMo 2 Furious
- OLMoTrace: Tracing Language Model Outputs Back to Trillions of Training Tokens
- Olmo 3
- Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding
- Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (full text)
- MolmoWeb: Open Visual Web Agent and Open Data for the Open Web
- MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction
- MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction (project page)
- NSF OMAI | Ai2