DeepSeek V4 Pro fingerprint change is just a cluster swap: weights bit-identical, likely Ascend backend
teortaxesTex · x · 2026-11-16
DeepSeek V4 Pro's model fingerprint changed (a307abda… → d400d062…), but community analysis shows it's not a model update: logprobs are bit-identical to the previous weights, the new backend is actually faster (p50 817ms vs 1394ms), and speed remains 90-100 tps. Testing on a sensitive question at fingerprint-change time matched V4 Pro exactly, suggesting the change is just a new cluster — possibly Ascend hardware. The same pattern held for DeepSeek Flash earlier.
More from Infra
- Rumor: DeepSeek V5 in pretraining, ~2T-param sparse MoE trained entirely on Huawei Ascend, no NVIDIA — DeryaTR_ · 2026-11-16
- Microsoft email outage delayed chlorine alerts at German theme park hotel, 15 guests harmed — evilsocket · 2026-11-16
- How Lisa Su took AMD from near-bankruptcy to $1T — and she's Jensen Huang's cousin — bigaiguy · 2026-11-16
- tandem.cpp v0.3.0: 2-6x Faster Prompt Processing, DeepSeek-V4.1 GGUFs Supported — smflx · 2026-11-16
- llmfit scans your hardware and tells you exactly which local LLMs will actually run — thisdudelikesAI · 2026-11-16
- SemiAnalysis reportedly cuts 2028 AI accelerator capex forecast by 28% — firstadopter · 2026-11-16