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.
Claude 3.5 Haiku, developed by Anthropic, offers a 200,000-token context window. Pricing is set at $1 per million input tokens and $5 per million output tokens, with potential savings of up to 90% through prompt caching and 50% via the Message Batches API. Released on November 4, 2024, this model excels in code completion, interactive chatbots, data extraction and labeling, as well as real-time content moderation.
| DeepSeek-R1 | Claude 3.5 Haiku | |
|---|---|---|
Web Site
| ||
Provider
| ||
Chat
| ||
Release Date
| ||
Modalities
| text | text |
API Providers
| DeepSeek, HuggingFace | Anthropic, AWS Bedrock, Vertex AI |
Knowledge Cut-off Date
| Unknown | 01.04.2024 |
Open Source
| Yes | No |
Pricing Input
| $0.55 per million tokens | $0.80 per million tokens |
Pricing Output
| $2.19 per million tokens | $4.00 |
MMLU
| 90.8% Pass@1 Source | Not available |
MMLU-Pro
| 84% EM Source | 65% 0-shot CoT Source |
MMMU
| - | Not available |
HellaSwag
| - | Not available |
HumanEval
| - | 88.1% 0-shot Source |
MATH
| - | 69.4% 0-shot CoT Source |
GPQA
| 71.5% Pass@1 Source | Not available |
IFEval
| 83.3% Prompt Strict Source | Not available |
SimpleQA
| - | - |
AIME 2024 | - | - |
AIME 2025 | - | - |
Aider Polyglot
| - | - |
LiveCodeBench v5
| - | - |
Global MMLU (Lite)
| - | - |
MathVista
| - | - |
Mobile Application |
Compare AI. Test. Benchmarks. Mobile Apps Chatbots, Sketch
Copyright © 2025 All Right Reserved.