# DeepSeek-R1-0528-Qwen3-8B

> An 8B reasoning model that distills DeepSeek-R1-0528's chain-of-thought into the Qwen3-8B base.

- Maker: DeepSeek
- Line: DeepSeek R1 Distill
- Canonical page: https://ai-tldr.dev/models/deepseek-r1-0528-qwen3-8b/
- Updated: 2026-06-18

## In short

DeepSeek-R1-0528-Qwen3-8B is an 8-billion-parameter open-weight reasoning model from DeepSeek, released on 29 May 2025 alongside the larger DeepSeek-R1-0528 update. It has a 131K-token context window. You can use it through OpenRouter.

## Facts

| Fact | Value | Source |
| --- | --- | --- |
| Context window | 131K | [1] |
| Licence | MIT · open weights | [1] |
| Released | 29 May 2025 | [2] |
| Parameters | 8B | [1] |

## Sources

1. [Model card — Hugging Face](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B)
2. [DeepSeek-R1-0528 release notes](https://api-docs.deepseek.com/news/news250528)

## Overview

DeepSeek-R1-0528-Qwen3-8B is an 8-billion-parameter open-weight reasoning model from DeepSeek, released on 29 May 2025 alongside the larger DeepSeek-R1-0528 update. DeepSeek built it by distilling the chain-of-thought reasoning from R1-0528 into Qwen3-8B Base — in other words, it takes the reasoning traces produced by the big 685B R1-0528 model and uses them to post-train Alibaba's small Qwen3-8B base model. The result is a compact model that thinks step by step before answering.

Its network architecture is identical to Qwen3-8B, so it runs with the same inference setup, but it ships with DeepSeek's own tokenizer configuration rather than Qwen's. DeepSeek recommends a sampling temperature of 0.6, supports system prompts, and (unlike some earlier R1 distills) does not require you to manually prepend a thinking token to trigger reasoning. The whole model is released under the permissive MIT License, which allows commercial use and further distillation.

DeepSeek positions DeepSeek-R1-0528-Qwen3-8B as a way to bring R1-grade reasoning to commodity hardware. On its own model card, DeepSeek reports that it reaches state-of-the-art results among open-source models on AIME 2024 — beating the original Qwen3-8B by 10 points and roughly matching the much larger Qwen3-235B-thinking on that math test.

## Benchmarks

- AIME 2024: 86% ([source](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B))
- AIME 2025: 76.3% ([source](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B))
- HMMT Feb 2025: 61.5% ([source](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B))
- GPQA Diamond: 61.1% ([source](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B))
- LiveCodeBench (2408-2505): 60.5% ([source](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B))

## Strengths

- State-of-the-art math reasoning for an 8B open model — 86.0 on AIME 2024 per DeepSeek's model card, +10 points over Qwen3-8B.
- Small enough to run locally on a single consumer GPU (or quantized on a laptop), bringing R1-style chain-of-thought to commodity hardware.
- Fully open under the MIT License, with weights on Hugging Face, so it can be self-hosted, fine-tuned, and distilled commercially.
- Drop-in Qwen3-8B inference compatibility (same architecture), with system-prompt support and no manual think-token needed.

## Best for

- Local and on-device reasoning where a hosted frontier model is impractical or privacy-sensitive.
- Math and competition-style problem solving (AIME / HMMT) on a small footprint.
- Cost-controlled coding and logic tasks via self-hosting under an open license.
- A base for further fine-tuning or distillation experiments under MIT.

## How to access

- [OpenRouter](https://openrouter.ai/deepseek/deepseek-r1-0528-qwen3-8b) — model id `deepseek/deepseek-r1-0528-qwen3-8b`

## DeepSeek R1 Distill versions

- DeepSeek-R1-0528-Qwen3-8B (29 May 2025): 8B distill of R1-0528 onto a Qwen3 base — this page
- [DeepSeek-R1-Distill-Llama-70B](https://ai-tldr.dev/models/deepseek-r1-distill-llama-70b/) (20 January 2025)
- [DeepSeek-R1-Distill-Qwen-32B](https://ai-tldr.dev/models/deepseek-r1-distill-qwen-32b/) (20 January 2025)
- [DeepSeek-R1-Distill-Qwen-14B](https://ai-tldr.dev/models/deepseek-r1-distill-qwen-14b/) (20 January 2025)
- [DeepSeek-R1-Distill-Llama-8B](https://ai-tldr.dev/models/deepseek-r1-distill-llama-8b/) (20 January 2025)
- [DeepSeek-R1-Distill-Qwen-7B](https://ai-tldr.dev/models/deepseek-r1-distill-qwen-7b/) (20 January 2025)
- [DeepSeek-R1-Distill-Qwen-1.5B](https://ai-tldr.dev/models/deepseek-r1-distill-qwen-1-5b/) (20 January 2025)

## FAQ

### What is DeepSeek-R1-0528-Qwen3-8B?

It is an 8-billion-parameter open-weight reasoning model that DeepSeek created by distilling the chain-of-thought from its larger DeepSeek-R1-0528 model into Alibaba's Qwen3-8B Base. It thinks step by step before answering and is released under the MIT License.

### How is it different from the original DeepSeek-R1-Distill-Qwen models?

The earlier January 2025 distills were built on Qwen 2.5 and Llama bases. This one, released 29 May 2025, distills the updated R1-0528 reasoning traces onto a newer Qwen3-8B base, which DeepSeek reports lifts AIME 2024 to 86.0 — state-of-the-art among open 8B models.

### Can I run DeepSeek-R1-0528-Qwen3-8B locally?

Yes. Its architecture is identical to Qwen3-8B, so it runs with standard Qwen3-8B inference tooling, and at 8B parameters it fits on a single consumer GPU (or quantized GGUF builds on a laptop). DeepSeek recommends a temperature of 0.6.

### What license does it use?

DeepSeek-R1-0528-Qwen3-8B is released under the MIT License, which permits commercial use and further model distillation.

## Links

- [Model card — Hugging Face](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B)
- [DeepSeek-R1-0528 release notes](https://api-docs.deepseek.com/news/news250528)
- [DeepSeek R1 — GitHub repository](https://github.com/deepseek-ai/DeepSeek-R1)
- [DeepSeek-R1-0528 — VentureBeat review](https://venturebeat.com/ai/deepseek-r1-0528-arrives-in-powerful-open-source-challenge-to-openai-o3-and-google-gemini-2-5-pro)
- [Run it locally — Unsloth guide](https://unsloth.ai/blog/deepseek-r1-0528)
- [Intelligence & price analysis — Artificial Analysis](https://artificialanalysis.ai/models/deepseek-r1-qwen3-8b)
- [API access — OpenRouter](https://openrouter.ai/deepseek/deepseek-r1-0528-qwen3-8b)

Source: AI/TLDR LLM registry, https://ai-tldr.dev/models/deepseek-r1-0528-qwen3-8b/