Open-source releaseModerate global significanceConfirmed confidence
DeepSeek releases DeepSeek V3.2
Another fine-tuning iteration on the V3 base, focused on lowering peak memory use and speeding up inference on local hardware.
What happened
Event details
DeepSeek V3.2 is a further fine-tuning iteration of the V3 base, optimizing peak VRAM usage and improving inference throughput in on-device settings. DeepSeek released the 671B open-source MoE model on December 1, 2025; the primary source is the official DeepSeek-V3.2 repository.
Assessment
Why it matters
Rated on the impact-v1 six-axis scale: technical 3/5, market 2/5, ecosystem 4/5, policy & social 1/5, scope 3/5, durability 3/5.
67/100Global significance score. Regional effects are recorded only when the evidence supports a meaningful difference.