Understand this story
On November 11, 2026, Google DeepMind released AlphaGenome 2, a conditional DNA sequence model that can predict the effects of DNA variants for specific tissues, cell types, or cell states. The model was trained on single-cell atlases covering 68 tissues and more than 107 million cells; on the TenK10K dataset, the proportion of causal variants identified increased from 16% to 30%, and directional gene effects rose from 23% to 43%. The model is available for non-commercial research through the AlphaGenome API, while commercial research is in private preview; AlphaGenome 2 Skills will subsequently be launched in Google Antigravity.
AI-generated from public reports
Latest development · Report
Google DeepMind releases AlphaGenome 2, which can predict which cell types DNA variants act onDevelopments
1 confirmed developments
- Source:Google DeepMind
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 12
36Kr · AI
Google DeepMind releases AlphaGenome 2, which can predict which cell types DNA variants act on
Google DeepMind has released AlphaGenome 2, a new-generation DNA sequence prediction model that can predict which cell types a DNA mutation acts in and which genes it affects, for use in tracing disease roots and finding drug targets. The model's training data covers 68 tissues and more than 107 million cells; in the TenK10K test covering 28 types of immune cells, its recall for identifying key causal variants is 30%, compared with 16% for the previous generation after fine-tuning on single-cell data, and its ability to determine whether a variant enhances or weakens gene activity rose from 23% to 43%. The model is now open to non-commercial research and can be called via the AlphaGenome API; it will later also launch on Google Antigravity in the form of AlphaGenome 2 Skills.
36Kr · AISelected
DeepMind Releases DNA Foundation Model AlphaGenome 2
Google DeepMind has released a new-generation DNA foundation model, AlphaGenome 2, and published a 78-page technical report, with the focus shifting to the non-coding regions that make up 98% of the human genome. The model uses single-cell data from 68 tissues and over 107 million cells; on the TenK10K benchmark, its pathogenic mutation identification rate rose from 16% to 30%, and its accuracy in determining whether gene activity is enhanced or weakened rose from 23% to 43%. The official release freely provides impact scores for nearly 7 million genetic variants across 510 cell states, and it plans to package the model as Skills integrated into Google Antigravity for use by Gemini 4.
The Decoder
Google DeepMind releases AlphaGenome 2, genomic AI shifts to single-cell-type resolution
Google DeepMind has released the genomic research model AlphaGenome 2, which can predict the effects of DNA sequences and their variants on specific human cell types, trained on single-cell atlases of 68 tissues and more than 107 million cells. On the TenK10K benchmark, which includes 28 primary immune cell types, it identified about 30% of causal variants and correctly determined whether a variant upregulates or downregulates a gene in 43% of cases, compared with 16% and 23%, respectively, for fine-tuned AlphaGenome 1. The model is now available via API for non-commercial research use, with commercial users going through a closed preview, and it has not been approved for clinical diagnosis.
heise online
Google DeepMind releases AlphaGenome 2, bringing cell state into DNA prediction
Google DeepMind has released AlphaGenome 2, a model that can predict how DNA changes affect gene activity across different cell types and cell states, analyzing segments up to 1 million base pairs long. In tests across 28 immune cell types, the new model identified 30% of suspected pathogenic DNA variants, compared with 16% for the previous generation. The model was trained on human and mouse data, with additional use of 107 million single-cell atlases from 68 human tissues; non-commercial research can use it for free via the AlphaGenome API (with usage limits), while commercial research is currently limited to a private preview, and the model has not been validated or approved for clinical or diagnostic use.
Testing Catalog
November 12 AI Daily Brief: Gemini Decision 1 Launches on API
Gemini Decision 1 is now available on the Gemini API and AI Studio, handling up to 1,000 questions per single call, with input priced at $0.10/1M tokens and no output charges. Claude Fable 5.2 has simultaneously landed on Google Cloud and is being rolled out in Copilot Cowork and Copilot Studio. Grok Bot 1.0 has been released, supporting shareable bots, Slack and Teams collaboration, and meeting invitations.
Zhidongxi
Google DeepMind releases AlphaGenome 2, using 107 million single-cell data to predict the impact of DNA variants
Google DeepMind has launched a new-generation DNA sequence model, AlphaGenome 2, which can predict how the same DNA sequence is read in different cell types, as well as the impact of a single base change on specific cells. The model was trained on more than 107 million single-cell data points, covering 68 tissues, and adds DNA methylation prediction. On the TenK10K dataset, the proportion of pathogenic causal variants it identified reached 30%, up from 16% for the previous generation; its accuracy in identifying the direction of gene expression changes was 43%, up from 23% for the previous generation. The model has been opened to non-commercial research through the AlphaGenome API, and commercial users can apply for a private preview version.
Phoenix Technology
Google DeepMind releases AlphaGenome 2, with predictions refined to the single-cell type
Google DeepMind has launched AlphaGenome 2, a new-generation genomic AI model that can predict the effects of DNA sequences and their variants on specific human cell types and states. The model was trained on over 107 million single-cell atlases from 68 tissues, covering human and mouse samples, and the number of gene expression-related variants it identifies is nearly twice that of the previous-generation model. DeepMind has opened the model to non-commercial research through the AlphaGenome API, while commercial research is in a limited preview phase.
Demis Hassabis
DeepMind releases AlphaGenome 2 DNA sequence model
Google DeepMind has released the AlphaGenome 2 DNA sequence model, which can predict the effects of DNA variants in specific cell types, for use in understanding complex diseases. The model is currently open to academic research via API. Demis Hassabis said that the same DNA sequence behaves differently in different cells, and that the model focuses on specific cells to help scientists understand the causes of disease.
SiliconANGLE
Google DeepMind Releases Gemini Robotics 2.5 and AlphaGenome 2
Google DeepMind has released Gemini Robotics 2.5, a new-generation robot vision-language-action model, and simultaneously launched the gene prediction model AlphaGenome 2. Gemini Robotics 2.5 completed automotive parts packaging with a 96% success rate in a real-world proof-of-concept test at Mercedes-Benz, with task success rates ranging from 30% to 100%, and was 20% to 30% faster than the previous-generation baseline; it was launched the same day through partners including Agile Robots, Apptronik, and Boston Dynamics. AlphaGenome 2 adds DNA methylation prediction and finer cell type and state information, and can adapt directly to new single-cell data without retraining, but it is intended only to guide research and has not been approved for clinical or diagnostic use.
Sundar Pichai
Google DeepMind releases AlphaGenome 2, which can predict the impact of DNA variants by cell type
Google DeepMind has released AlphaGenome 2, a model that can predict how different cell types interpret DNA, helping researchers connect genetic findings with the root causes of disease in specific cells. Built on AlphaGenome 1, it can predict the impact of DNA variants on specific cell types, and is used to reveal the cellular-level drivers of disease.
cnBeta
DeepMind releases AlphaGenome 2, refining the analysis of genetic variant effects down to single-cell types
Google DeepMind has released AlphaGenome 2, refining the granularity of its analysis of genetic variants' effects on gene activity from the tissue level down to specific cell types and cell states. The model was trained on more than 100 million single-cell data points, covering 68 human tissues, and can identify roughly twice as many genetic variants that may alter cellular gene expression as previous state-of-the-art models, and can make predictions for cell types and states not directly present in training. AlphaGenome 2 will be opened to non-commercial researchers through the AlphaGenome API starting November 11, with commercial organizations having only limited private preview access; it will not replace the previous-generation model, which remains in use for large-scale genome-wide predictions.
Nov 11
Unite.AI
Google DeepMind releases AlphaGenome 2, predicting the effects of DNA variants by cell state
Google DeepMind released AlphaGenome 2 on November 11, 2026. This deep learning model can predict how DNA variants affect gene regulation in specific cell types and states. Its training data includes more than 100 million single-cell atlases. It is open for non-commercial research and offers a private preview for commercial research. On TenK10K, a single-cell immune dataset covering 28 cell types, it recovered 30% of putative causal expression-associated variants at 90% precision, compared with 16% for the fine-tuned predecessor; the effect-direction recovery rate was 43% versus 23%. The model evaluates only one condition at a time, and its running overhead is higher than that of multi-task models with fixed outputs. The paper also notes that it can recover only a small number of candidate variants, and that the cell-state representations come from observational steady-state data, reflecting correlation rather than causation.
Google DeepMindFirst-partySelected
Google DeepMind releases AlphaGenome 2, predicting the impact of DNA variants by cell type
Google DeepMind has released AlphaGenome 2, a conditional DNA sequence model that can predict the impact of DNA variants for specific tissues, cell types, or cell states. The model was trained on single-cell atlases covering 68 tissues and more than 107 million cells. On the TenK10K dataset, the proportion of causal variants identified rose from 16% to 30%, and directional gene effects rose from 23% to 43%. The model is available for non-commercial research through the AlphaGenome API; commercial research is in private preview, and AlphaGenome 2 Skills will later be launched in Google Antigravity.
Heat trend for this story
- Current
- 5
- Period peak
- 81 Nov 12 21:00
- 24-hour change
- -89%
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.