DeepSeek-R1 is a 671B parameter Mixture-of-Experts (MoE) model with 37B activated parameters per token, trained via large-scale reinforcement learning with a focus on reasoning capabilities. It incorporates two RL stages for discovering improved reasoning patterns and aligning with human preferences, along with two SFT stages for seeding reasoning and non-reasoning capabilities. The model achieves performance comparable to OpenAI-o1 across math, code, and reasoning tasks.
| DeepSeek-R1 | Qwen 3 | |
|---|---|---|
Provider | ||
Web Site | ||
Release Date | Jan 21, 2025 3 months ago | Apr 29, 2025 16 hours ago |
Modalities | text | - |
API Providers | DeepSeek, HuggingFace | - |
Knowledge Cut-off Date | Unknown | - |
Open Source | Yes | Yes (Source) |
Pricing Input | $0.55 per million tokens | - |
Pricing Output | $2.19 per million tokens | - |
MMLU | 90.8% Pass@1 Source | - |
MMLU Pro | 84% EM Source | - |
MMMU | - | - |
HellaSwag | - | - |
HumanEval | - | - |
MATH | - | - |
GPQA | 71.5% Pass@1 Source | - |
IFEval | 83.3% Prompt Strict Source | - |
Mobile Application | - |
Compare AI. Test. Benchmarks. Mobile Apps Chatbots, Sketch
Copyright © 2025 All Right Reserved.