DeepSeek / DeepSeek V3.1
DeepSeek V3.1 Distillation
Estimated until measured during engineering
Before → After
Base releaseVerified checkpoint
671.000B
671.000B
DistillationTeacher / Student
Quality retention target
Quality retention target
Production artifactArtifact
Lineage and benchmark report
Lineage and benchmark report
GPU targeting
Weights and KV cacheEstimated until measured during engineering
Estimated VRAMMeasured and certified results are delivered with the project.
Runtime packagingtransformers