Hello CM4AI DREAM Challenge Participants! Thank you to everyone who has already submitted their models. We will conduct the first round of optional interim evaluations starting in December. This will include evaluating your method on several thousand samples across hundreds of proteins. The first interim evaluation will use submissions received by 11:59 pm PT on Tuesday, December 1, 2026. Final submissions will be due Monday, February 1, 2027. Participation in the interim evaluation is optional and is not a requirement for the final evaluation. To manage resources, we are limiting the number of submissions evaluated per team; this initial round is your first opportunity to benchmark your method's performance. Submitting to this phase does not lock in your final entry; you may use these interim benchmark scores to refine your model before the final submission deadline. For this round, we will evaluate only one model per individual or team; specifically, the most recent submission received by the deadline. Before each evaluation, we will run a validation run  to verify they are functioning within our benchmarking system. If a submission does not pass validation, we will allow teams to provide a replacement submission that must be received by December 8th to be part of the phase 1 evaluation.  As a reminder, all evaluations are performed using a test set of KOLF2.1J cells. We have provided a 'fine-tuning' set of KOLF data as part of the training set to allow for model refinement. You may also use the fine-tuning set to develop your own internal validation benchmarks. The evaluation code for this challenge can be found at https://github.com/cm4ai-anvil-dream-challenge/cm4ai-evaluate.  As always, all communications are through Synapse, and the CM4AI DREAM Challenge Discussion forum is the best way to ask questions.  Best regards, The CM4AI DREAM Challenge Team Emma Lundberg Kyle Ellrott Shurjo Sen