Skip to main contentThe Local Bench

Math and reasoning · DeepSeek

DeepSeek R1 Distill Qwen 7B

Hugging Face: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B · checked 2026-10-13

The original R1 distill at 7B; superseded by R1-0528 Qwen3 8B but widely packaged.

Datasheet

Maker
DeepSeek
Parameters
7.62 BVerifiedsafetensors.total from the HF API Source
Licence
MITVerified
Access
Public download
Task (HF pipeline)
text-generation
First published on HF
2025-01-20
Last updated on HF
2025-02-24
Weights in memory
15 GB at 16-bit · 4.6 GB at ~4.85 bitsEstimateParameters × bits ÷ 8. Excludes context cache and runtime overhead; not every model has a 4-bit build.

Benchmark

55.5%
AIME 2024, pass@1 · #7 of 7 scored in this category

Distilled-model evaluation table, model card. Source

Other published results

  • MATH-500, pass@1: 92.8%Sourced Source

See the full math and reasoning ranking

The description summarises the model card at https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B. Parameters and licence are read from the Hugging Face API; the weights-in-memory line is our arithmetic, not a measurement.