[whatmodel.sh](../../README.md) / DeepSeek R1 0528 Qwen3 8B # DeepSeek R1 0528 Qwen3 8B DeepSeek · released 2025-05 · 8.19B params · mit **score 57.2** ±0.9 · #242 of 732 · from 20 benchmark results across 5 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at ● marks DeepSeek R1 0528 Qwen3 8B among every model on that task or test, worst to best. 
task         vs the others          rank  counts  tests
agent tasks  ▁▁▃●█▄▅▅▄▄▂▁   #273 of 313*    5.7%      1
coding       ▁▁▁▃▅▇●▇▅▄▃▁    #259 of 469    8.8%      3
math         ▁▁▃▅█▆●█▄▂▁▁    #156 of 431     45%      6
agent tasks · #273 of 313 · 5.7% of its score Trails [Granite 4.0 H 1B](granite-4-0-h-1b.md), [Granite 4.0 1B](granite-4-0-1b.md), [SmolLM3 3B](smollm3-3b.md) here, though they rank lower overall. 
test                 result  vs the others in its run  difficulty            counts  why
Terminal-Bench Hard    1.4%  ●▅▆▂▂▆▂▂█▅▅▂ beats 0/27   very hard · avg 1.3%    5.7%  sharp · too hard for it
coding · #259 of 469 · 8.8% of its score Beats [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md), [Solar Open 100B (102B-A12B)](solar-open-100b.md), [GLM-4.7-Flash](glm-4-7-flash.md) here, though they rank higher overall. Trails [DeepSeek LLM 67B Chat](deepseek-llm-67b-chat.md), [QwQ-32B-Preview](qwq-32b-preview.md), [InternLM2.5 20B Chat](internlm2-5-20b-chat.md) here, though they rank lower overall. 
test                                result  vs the others in its run  difficulty            counts  why
LiveCodeBench v6 (window unspecif…   57.2%  █·●·▄····▄·▄ beats 2/5    medium · avg 41%        5.7%  at its level
LiveCodeBench v6 (Aug 2024–May 20…   60.5%  ▂●▅▃█▃▂····▂ beats 2/17   medium · avg 48%        1.9%  at its level · small run
SciCode main problems                 2.6%  ●··▄····█▄·▄ beats 0/5    very hard · avg 3.8%    1.2%  too hard for it
math · #156 of 431 · 45% of its score Beats [MiMo-V2.5 (310B-A15B)](mimo-v2-5.md), [MiniMax-M2.7](minimax-m2-7.md), [MiniMax-M2.1](minimax-m2-1.md) here, though they rank higher overall. Trails [Qwen3-VL 4B Thinking](qwen3-vl-4b-thinking.md), [Nemotron 3 Nano 30B-A3B](nemotron-3-nano-30b-a3b.md) here, though they rank lower overall. 
test                             result  vs the others in its run   difficulty            counts  why
AIME 2024                         83.3%  █▄···●·····▄ beats 4/5     medium · avg 62%         13%  sharp · several runs
AIME 2025                         75.8%  ▄·█··●····▄▄ beats 3/5     medium · avg 52%         13%  sharp · at its level · several runs
HMMT 2025 (edition unspecified)   54.3%  ▄··▄▄·●····█ beats 3/5     hard · avg 29%          9.1%  sharp · at its level
OTIS mock AIME 2024–2025          43.9%  ▃▃▄▃▃●▇▇█▃▂▅ beats 73/197  very hard · avg 7.2%    5.4%  sharp · at its level · noisy
MATH-500                          96.8%  ▄·▄·····●··█ beats 3/5     very easy · avg 92%     3.7%  maxed out
HMMT February 2025 (maker)        61.5%  ▁▁▁▃▁●▅▃██▁▄ beats 6/31    medium · avg 61%           —  not counted yet: no other model in the …
other tests (8) 
test                                result  vs the others in its run   difficulty            counts  why
GPQA Diamond                         61.4%  ●·········▂▂ beats 3/5     hard · avg 44%           21%  sharp · at its level · several runs
MMLU-Pro                             69.1%  █··●··███··█ beats 1/5     medium · avg 57%        5.4%  at its level
BRIDGE clinical text, zero-shot a…   36.9%  ▁▁▁▁▁▂▂▃▄●█▃ beats 54/97   hard · avg 31%          4.9%  blunt · at its level
Humanity's Last Exam                  5.6%  █··●█···█·██ beats 1/5     very hard · avg 3.7%    4.8%  too hard for it
IFBench (aggregation unspecified)    29.2%  ●██···██···█ beats 0/5     hard · avg 34%          1.5%  at its level
EuroEval HellaSwag (English), tes…   69.0%  ▃·▁▁▂▁▂▅●▇▄▁ beats 48/100  medium · avg 67%        1.2%  sharp · at its level · noisy
Chess Puzzles (Epoch)                 3.0%  ●▁▁▁▁▂▃▃▃▂▁▁ beats 24/137  very hard · avg 5.8%    0.9%  sharp · too hard for it · noisy
EuroEval Life in the UK (English …   83.0%  ▁▁▁▁▁▁▁▂▄●▄▃ beats 85/146  easy · avg 82%          0.6%  noisy
## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its *counts* column above; here it is added up by who ran the tests. [Method](../method.md) 
source                                    results  share of the score
Falcon-H1R: Pushing the Reasoning Front…       11   67% ▇▇▇▇▇▇▇▇
deepseek-ai/DeepSeek-R1-0528-Qwen3-8B m…        4   19% ▇▇
Epoch AI                                        2    6% ▇
BRIDGE Medical Leaderboard                      1    5% ▇
EuroEval leaderboards                           2    2% ▇
prior: pull toward the middle                        0% ▇
weighted average                                   = 57.2
Licences: Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scaling ([CC BY 4.0](../../licenses/report-2601.02346v1.txt)) · deepseek-ai/DeepSeek-R1-0528-Qwen3-8B model card ([MIT](../../licenses/card-deepseek-ai-deepseek-r1-0528-qwen3-8b.txt)) · Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE))
Not counted (2) 
source                  test                score  why
Epoch AI                GPQA Diamond         9.3%  failed run: 9.3% on a 4-option test (chance 25%) for a model that scores 61% elsewhere; a reasoning model scoring far below guessing points to unparsed or truncated answers, not ability.
deepseek-ai/DeepSeek-…  HMMT February 2025  61.5%  too few models have it yet
## How to run it ### Through an API api (Novita AI) · $0.068/M blended · $0.06 in / $0.09 out per 1M tokens · [novita.ai](https://novita.ai/models) [price source](https://github.com/BerriAI/litellm/blob/c19ce71bcdad234ca87a7af10cee45375ce811a8/model_prices_and_context_window.json): USD per 1M text tokens, standard uncached input and output. Model endpoint deepseek/deepseek-r1-0528-qwen3-8b. Product: Serverless chat completions. API model: deepseek/deepseek-r1-0528-qwen3-8b. Settings: service tier=standard, batch=no, cache=uncached input. Additional/excluded charges: Tools, cache storage, priority service and other modalities excluded.. Currency: USD. · [context source](https://github.com/BerriAI/litellm/blob/c19ce71bcdad234ca87a7af10cee45375ce811a8/model_prices_and_context_window.json): provider maximum input tokens; not an intrinsic model maximum or legacy max\_tokens ### llama.cpp (recommended) 1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md) ```sh # macOS / Linux curl -LsSf https://llama.app/install.sh | sh # Windows irm https://llama.app/install.ps1 | iex ``` 2. Run DeepSeek R1 0528 Qwen3 8B (Q4_K_M, needs ~5 GB, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF/tree/335076f2c9044728bd2d906527150757e03ba35e)) ```sh llama cli -hf lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_K_M ``` 3. Or start a local server and open the address it prints (http://localhost:9931) ```sh llama serve -hf lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_K_M ``` Other versions: Q8_0 (~8.7 GB) `llama cli -hf lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q8_0` ### LM Studio 1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download) 2. Run DeepSeek R1 0528 Qwen3 8B (Q4_K_M, needs ~5 GB, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF/tree/335076f2c9044728bd2d906527150757e03ba35e)) ```sh lms get https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF@Q4_K_M -y ``` 3. Then run lms ls to see its name, and lms chat followed by that name Other versions: Q8_0 (~8.7 GB) `lms get https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF@Q8_0 -y` · 4bit (~4.6 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-4bit -y` · 8bit (~8.7 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-8bit -y` ### Ollama 1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download) ```sh # Linux curl -fsSL https://ollama.com/install.sh | sh ``` 2. Run DeepSeek R1 0528 Qwen3 8B (Q4_K_M, needs ~5.41 GB, from [ollama.com](https://ollama.com/library/deepseek-r1/tags)) ```sh ollama run deepseek-r1:8b ``` Other versions: Q8_0 (~10.33 GB) `ollama run deepseek-r1:8b-0528-qwen3-q8_0` ### MLX (Mac) 1. Install MLX (Mac). Apple silicon Macs only, text only. ```sh # with uv uv tool install mlx-lm # or pip install mlx-lm ``` 2. Run DeepSeek R1 0528 Qwen3 8B (4bit, needs ~4.6 GB, Mac only, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-4bit/tree/1b6065ad0cd7256f662c9ef6fd0ed2dfd06e286f)) ```sh mlx_lm.chat --model lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-4bit ``` 3. Or start an OpenAI-compatible server at http://localhost:8080/v1 ```sh mlx_lm.server --model lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-4bit ``` Other versions: 8bit (~8.7 GB, Mac only) `mlx_lm.chat --model lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-8bit` Model repository and documentation: [huggingface.co/deepseek-ai](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B) ## Published speed measurements Reference results for the hardware, engine and settings below. These are not verified speeds of the downloads above or predictions for your machine. 
hardware                      precision               speed  engine and settings
Apple M3 Pro, 18 GB           q4_k_m            21.42 tok/s  ollama (version not recorded) · l…
NVIDIA GeForce GTX 1650 wit…  q4_k_m · layers…   3.58 tok/s  ollama (version not recorded) · l…
Each speed links to its source, which has the full settings. [interactive version](https://whatmodel.sh/model/deepseek-r1-0528-qwen3-8b) · [JSON](https://whatmodel.sh/api?format=json&all=1)