NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/R1 0528
DeepSeek logo

R1 0528

by DeepSeek · Released May 2025

Open Source
55.9
avg score
Rank #112
Compare
Better than 64% of all models
Context
164K tokens (~82 books)
Input $/1M
$0.50
Output $/1M
$2.15
Type
text
License
Open Source
Benchmarks
25 tested
Data as of
About

May 28th update to the original DeepSeek R1 Performance on par with OpenAI o1, but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active...

Tested on 25 benchmarks · BenchGecko score 55.9. Top scores: Chatbot Arena Elo — Overall (1421.6%), MATH level 5 (96.6%), OpenCompass — AIME2025 (89.0%).

Looking for similar performance at lower cost?
Qwen2.5 Coder 7B Instruct scores 56.3 (101% as good) at $0.03/1M input · 94% cheaper
Capabilities
coding
58.0
#54 globally
reasoning
33.5
#112 globally
math
73.6
#37 globally
knowledge
56.9
#82 globally
speed
24.0
#97 globally
language
79.2
#59 globally
Benchmark Scores
Compare All
Tested on 25 benchmarks · Ranked across 7 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
codingCompare coding →
Aider polyglot

Multi-language code editing from Aider. Tests editing ability across Python, JavaScript, TypeScript, Java, C++, Go, Rust, and more.

71.4·
OpenCompass — LiveCodeBenchV6

OpenCompass Live Code Bench v6. Fresh competitive programming problems to evaluate code generation without memorization.

61.0·
WeirdML

Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.

41.6·
reasoningCompare reasoning →
HELM — WildBench

Stanford HELM WildBench evaluation. Tests reasoning on challenging real-world tasks.

82.8·
SimpleBench

Deceptively simple questions that humans find easy but AI models often get wrong. Tests common sense and reasoning gaps.

29.0·
ARC-AGI

Abstraction and Reasoning Corpus. Tests fluid intelligence through novel visual pattern recognition puzzles. Core measure of general intelligence.

21.2·
mathCompare math →
MATH level 5

Competition-level math from AMC, AIME, and olympiad problems. Level 5 is the hardest tier, requiring creative problem-solving.

96.6·
OpenCompass — AIME2025

OpenCompass evaluation on AIME 2025 problems. Tests mathematical reasoning on fresh competition problems.

89.0·
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

66.4·
Quick compare:OpenAI logovs GPT-5 Minivs Dolphin 2.9.1 Yi 1.5 34bAlibaba logovs Qwen2.5 Coder 7B Instruct
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Recently Happened
R1 0528 pricing increased 11%
Apr 13, 2026
Similar Models
OpenAI logo
GPT-5 Mini
OpenAI
56.0$0.25/1M
Dolphin 2.9.1 Yi 1.5 34b
DPHN
55.6n/a
Alibaba logo
Qwen2.5 Coder 7B Instruct
Alibaba
56.3n/a
Links
Info
DeepSeekPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@DeepSeek
Source Code
GitHubHugging Face
BenchGecko API
deepseek-r1-0528
Specifications
  • Typetext
  • Context164K tokens (~82 books)
  • ReleasedMay 2025
  • LicenseOpen Source
  • StatusActive
  • Cost / Message~$0.003
Available On
DeepSeek logoDeepSeek$0.50
Categories
codingreasoningmathknowledgespeedlanguage
Learn More
context-windowtransformeropen weightstokens
Share & Export
Tweet
Related Models
OpenAI logoGPT-5 Mini
56.0
Dolphin 2.9.1 Yi 1.5 34b
55.6
Alibaba logoQwen2.5 Coder 7B Instruct
56.3
Alibaba Qwen logoQwen2.5 Coder 7B Instruct
56.3
R1 0528 is an open-source text AI model by DeepSeek, released in May 2025. It has an average benchmark score of 55.9. Context window: 164K tokens.

Related Models

GPT-5 Mini · OpenAIDolphin 2.9.1 Yi 1.5 34b · DPHNQwen2.5 Coder 7B Instruct · AlibabaQwen2.5 Coder 7B Instruct · Alibaba QwenGrok 4.20 · xAI

Benchmarks

Chatbot Arena Elo — OverallMATH level 5OpenCompass — AIME2025OpenCompass — MMLU-ProHELM — WildBench

Related Pages

DeepSeek · ProviderDeepSeek · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary

Key facts · as of 2026-10-05

  • R1 0528 by DeepSeek. BenchGecko score 55.9, rank 112 of 312 scored models (normalized average of public benchmark scores).
  • List price $0.50 input · $2.15 output per 1M tokens (as of 2026-10-05).
  • Sold by 4 providers (as of 2026-10-05): DeepInfra (fp4) $0.50 in / $2.15 out · SiliconFlow (fp8) $0.50 in / $2.18 out · StreamLake $0.57 in / $2.29 out · Novita (fp8) $0.70 in / $2.50 out. Every provider

How to cite · data as of 2026-10-05

R1 0528 · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/deepseek-r1-0528

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP