DeepSeek-R1-Distill-Qwen-7B

Name: DeepSeek-R1-Distill-Qwen-7B
Author: DeepSeek

DeepSeek · DeepSeek-R1 distill · released 2025-01-20 · data checked 2026-11-01 Needs verification

DeepSeek-R1-Distill-Qwen-7B is a 7.6B-parameter dense model from DeepSeek (China), released 2025-01-20 under the MIT license. It scores 39.8 on the Small LM Index, #19 of 26 small models. It needs about 5.7 GB of memory at 4-bit with an 8K-token context.

Small LM Index

39.8

#19 of 26 scored small models

Likely 35.1–44.7 · from 3 of 6 core benchmarks

Parameters

7.6B

Dense

Price per 1M tokens

—

No hosted API on OpenRouter; self-host

Context window

128K

tokens

Benchmarks

Each bar shows this model's score. The tick marks the best small model on that benchmark. Sources: 3 reported by DeepSeek. How we label sources.

GPQA Diamond Lab-reported Knowledge49.1

Class best: Qwen3.5 9B 81.7

AIME Lab-reported Math · AIME 202455.5

Class best: Apriel 1.6 15B Thinker 88.0

LiveCodeBench Lab-reported Coding37.6

Class best: Granite 4.2 8B 73.2

Not reported: estimated

What the index model expects from this model's other scores. Not a measurement.

MMLU-Pro≈ 65

Humanity's Last Exam≈ 4

SWE-bench Verified≈ 20

Specifications

Lab: DeepSeek (CN · China)
Architecture: Dense transformer
Total parameters: 7.6B
Active parameters: All (dense)
Size class: Small
Context window: 128K tokens
Input modalities: text
Reasoning: Reasoning (thinking)
License: MIT
Openness: Open weights
Memory to run: ~5.7 GB at 4-bit, 8K context (more sizes)
Hugging Face: 321.5K downloads (30 days) · 950 likes

Weights on Hugging Face ↗

Compare with

DeepSeek-R1-Distill-Qwen-7B vs Falcon H1R 7B (7.6B, index 57.5)
DeepSeek-R1-Distill-Qwen-7B vs Hunyuan 7B Instruct (7.5B, index 52.5)
DeepSeek-R1-Distill-Qwen-7B vs OLMo 3 7B Think (7.3B, index 43.4)
DeepSeek-R1-Distill-Qwen-7B vs MiMo-7B-RL (7.8B, index 40.7)
DeepSeek-R1-Distill-Qwen-7B vs Gemma 3n E4B (7.8B, index 18.2)

Sources

https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B

Something wrong here? Report it.

Run it locally

Uses unsloth/DeepSeek-R1-Distill-Qwen-7B-GGUF (Q4_K_M). Swap the tag for another size: F16, Q2_K, Q2_K_L, Q3_K_M, Q4_K_M, Q5_K_M, Q6_K, Q8_0.

Ollama

ollama run hf.co/unsloth/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M

llama.cpp

llama-server -hf unsloth/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M

Then open http://localhost:8080 for a chat window.

MLX (Mac)

pip install -U mlx-lm
mlx_lm.chat --model mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bit

vLLM (GPU server)

vllm serve deepseek-ai/DeepSeek-R1-Distill-Qwen-7B

Full-precision weights, for a server GPU.

Open in LM Studio

Memory needed: weights, context cache and runtime buffers.

Quantization	4K context	32K context	128K context
16-bit F16	16 GB	18 GB	23 GB
8-bit Q8_0	9.0 GB	11 GB	16 GB
4-bit Q4_K_M	5.5 GB	7.1 GB	13 GB

Check it against your GPU, Mac or phone →