Unverified50% confidenceReleaseExact time
DeepSeek released the 763B-parameter model weights on Hugging Face under the MIT license in FP8 format, with 13 quantized versions and 2 fine-tunes already available.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Related Entities
Related Claims
UnverifiedDeepGEMM is a high-performance matrix multiplication library designed by DeepSeek specifically for its MoE models, supporting low-precision formats such as bf16 and fp8.64% similarUnverifiedDeepSeek-Math 7B is an open-source mathematics large model with 7B parameters, whose RL version with tool invocation can achieve approximately 60% accuracy on the MATH benchmark, approaching GPT-4 level.64% similarUnverifiedThe DeepSeek reverse-prompt model offers two versions: 1B and 7B.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1054799API
curl https://kongchang.com/api/v1/knowledge/claims/1054799MCP
get_claim(id=1054799)