whatmodel.sh / DeepSeek R1 0528 Qwen3 8B

DeepSeek R1 0528 Qwen3 8B

+ contribute

DeepSeek R1 0528 Qwen3 8B is an open-weights language model from DeepSeek, released May 2025. Its published size is 8.19B parameters and a 128k-token context window. Licence: mit.

57.3±0.9#230 of 719 LLMsfrom 20 benchmark results across 5 sources

$ how it compares

The LLMs ranked either side of DeepSeek R1 0528 Qwen3 8B, out of 828 with a score, then more from DeepSeek.

#	model	score	where	RAM	released
273	GPT-4o (2024-08-06) OpenAI	57±1	api $4.38/M		2024-08
274	Gemini 2.0 Flash Thinking Experimental (01-21) Google	57±2	comparison only		2025-01
275	DeepSeek R1 0528 Qwen3 8B DeepSeek	57±1	api $0.068/M +local	~6.3GB	2025-05
276	Granite 4.2 3B IBM	57±1	api $0.052/M +local	~3.7GB	2026-08
277	Command A (03-2025) Cohere	57±1	api $4.38/M +local	~37GB	2025-03
	more from DeepSeek
21	DeepSeek V4.1 Flash DeepSeek	78±1	api $0.26/M +local	~537GB	2026-09
31	DeepSeek V4 Pro (0813) DeepSeek	76±1	api $1.63/M +local	~1010GB	2026-08
35	DeepSeek V3.2 Speciale DeepSeek	76±3	comparison only

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #262 of 300 * · 5.9% of its score · 1 test

Trails Granite 4.0 H 1B, Granite 4.0 1B, Qwen2.5 3B Instruct 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 1.4%	▇▇ 5.9%	sharp · too hard for it

coding▁▁▁▃▆██▇▆▄▃▁ #248 of 458 · 8.9% of its score · 3 tests

Beats Qwen3-VL 32B Instruct, Solar Open 100B (102B-A12B), GLM-4.7-Flash here, though they rank higher overall. Trails QwQ-32B-Preview, InternLM2.5 20B Chat, Mistral Small Instruct 22B (2409) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
LiveCodeBench v6 (window unspecified)	57.2% ✎	█·▄·▄····▄·▄ beats 2/5	medium · avg 42%	▇▇ 5.5%	at its level
LiveCodeBench v6 (Aug 2024–May 2025) lab	60.5% ✎	▂▅▅▃█▃▂····▂ beats 2/17	medium · avg 48%	▇ 2.2%	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▁▂▃▅▆▅▄█▄▁▁▁ #153 of 398 · 48% of its score · 6 tests

Beats MiMo-V2.5 (310B-A15B), MiniMax-M2.7, MiniMax-M2.1 here, though they rank higher overall. Trails Granite 4.2 3B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
AIME 2024 +1	83.3% ✎	█▄···█·····▄ beats 4/5	medium · avg 62%	▇▇▇▇▇▇ 14%	sharp · several runs
AIME 2025 +1	75.8% ✎	▄·█··▄····▄▄ beats 3/5	medium · avg 53%	▇▇▇▇▇ 13%	sharp · at its level · several runs
HMMT 2025 (edition unspecified)	54.3% ✎	▄··▄▄·▄····█ beats 3/5	hard · avg 29%	▇▇▇▇ 10%	sharp · at its level
OTIS mock AIME 2024–2025	43.9% ✎	▃▃▄▃▃▃▇▇█▃▂▅ beats 73/193	very hard · avg 8.5%	▇▇▇ 6.4%	sharp · at its level · noisy
MATH-500	96.8% ✎	▄·▄·····█··█ beats 3/5	very easy · avg 92%	▇▇ 4.3%	maxed out
HMMT February 2025 lab	61.5% ✎	▁▁▁▃▁▅▅▃██▁▄ beats 6/31	medium · avg 58%	—	not counted yet: no other model in the same run to compare it with

other tests8 not in a task · 37% of its score

test	result	vs the others in its run	difficulty	counts	why
GPQA Diamond +2	61.4% ✎	█·········▂▂ beats 3/5	hard · avg 45%	▇▇▇▇▇▇▇▇ 19%	sharp · at its level · several runs
MMLU-Pro	69.1% ✎	█··█··███··█ beats 1/5	medium · avg 57%	▇▇ 5.2%	at its level
Humanity's Last Exam	5.6% ✎	█··██···█·██ beats 1/5	very hard · avg 3.8%	▇▇ 4.7%	too hard for it
BRIDGE clinical text, zero-shot average (87 tasks)	36.9% ✎	▁▁▁▁▁▂▂▃▄▆█▃ beats 54/97	hard · avg 32%	▇▇ 4.0%	blunt · at its level
IFBench (aggregation unspecified)	29.2% ✎	███···██···█ beats 0/5	hard · avg 35%	▇ 1.5%	at its level
EuroEval HellaSwag (English), test split, 5-shot generative MC, accuracy	69.0% ✎	▃·▁▁▂▁▂▅█▇▄▁ beats 48/100	medium · avg 68%	▇ 1.0%	at its level · noisy
Chess Puzzles (Epoch)	3.0% ✎	█▁▁▁▁▂▁▄▂▂▁▁ beats 24/135	very hard · avg 5.9%	▇ 0.8%	sharp · too hard for it · noisy
EuroEval Life in the UK (English knowledge MC), test split, 5-shot, accuracy	83.0% ✎	▁▁▁▁▁▁▁▂▄█▄▃ beats 85/146	easy · avg 83%	▇ 0.5%	noisy

$ where to run it

Through an API

api · Novita AI · $0.068/M · $0.06 in / $0.09 out per 1M tokens · 128k context · novita.ai ↗ · price sourceUSD 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.

llama.cpp · recommended

version

Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
Windows
irm https://llama.app/install.ps1 | iex
other ways to install ↗
Run DeepSeek R1 0528 Qwen3 8B · needs ~5 GB · from huggingface.co/lmstudio-community
chat in the terminal
llama cli -hf lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_K_M

LM Studio

version

Install LM Studio · Open the app once before using the lms command.
download it (macOS / Windows / Linux) ↗
Run DeepSeek R1 0528 Qwen3 8B · needs ~5 GB · from huggingface.co/lmstudio-community
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-GGUF@Q4_K_M -y
Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

Install Ollama
download it (macOS / Windows) ↗
Linux
curl -fsSL https://ollama.com/install.sh | sh
Run DeepSeek R1 0528 Qwen3 8B · needs ~5.41 GB · from ollama.com
chat in the terminal
ollama run deepseek-r1:8b

MLX (Mac)

version

Install MLX (Mac) · Apple silicon Macs only, text only.
with uv
uv tool install mlx-lm
or
pip install mlx-lm
Run DeepSeek R1 0528 Qwen3 8B · needs ~4.6 GB, Mac only · from huggingface.co/lmstudio-community
chat in the terminal
mlx_lm.chat --model lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-4bit
Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model lmstudio-community/DeepSeek-R1-0528-Qwen3-8B-MLX-4bit

Weights and documentation: huggingface.co/deepseek-ai ↗

$ how the score was worked out · Falcon-H1R: Pushing the Reaso… 67%, deepseek-ai/DeepSeek-R1-0528-… 20%, Epoch AI — internal evaluatio… 7% and 2 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The 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's added up by who ran the tests. How it works

source	results	share of the score	licence
Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scaling	11	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 67%	CC BY 4.0
deepseek-ai/DeepSeek-R1-0528-Qwen3-8B model card · lab-reported	4	▇▇▇▇▇▇ 20%	MIT
Epoch AI — internal evaluations	2	▇▇ 7%	CC BY 4.0
BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)	1	▇ 4%	MIT
EuroEval leaderboards (results collection), English generative leaderboard	2	▇ 2%	MIT
prior · pull toward the middle		▇ 0%
weighted average		= 57.3

not counted (2)

Epoch AI — internal evaluations	GPQA Diamond [correct]	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-R1-0528-Qwen3-8B model card	HMMT February 2025 [correct]	61.5%	fewer than three models have this test yet, or this source ran it for this model alone

$ improve this page

Update model information Add a place to run it Add a test result Propose a benchmark Correct a benchmark score Edit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated 2026-10-11 02:20 UTC · sources and licences · json