Fact

DeepSeek — Model Parameters: 671 billion

confirmed99% confidence

1 evidence check

Last checked: 3/31/2026

The source text directly confirms both key specifications of the claim: (1) DeepSeek-V3 has 671B total parameters, and (2) 37B parameters are activated per token. The abstract explicitly states these figures, and the document is the official DeepSeek-V3 Technical Report submitted in December 2024, matching the temporal context of the claim (2024-12). The MoE architecture is also confirmed.

Evidence — 1 source, 1 check

arxiv.org/abs/2412.19437(1 check)

confirmed99%primaryHaiku 4.5 · 3/31/2026

Found: DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token

Note: The source text directly confirms both key specifications of the claim: (1) DeepSeek-V3 has 671B total parameters, and (2) 37B parameters are activated per token. The abstract explicitly states these figures, and the document is the official DeepSeek-V3 Technical Report submitted in December 2024, matching the temporal context of the claim (2024-12). The MoE architecture is also confirmed.

Debug info

Record type: fact

Record ID: f_qbe0bnJl8Q