Deepseek: An Open Source AI Model that outperformed: - Llama 3.2 -...
CJ Zafir@cjzafir
53 views
Jan 20, 2025
~3 min read
Advertisement
1
Deepseek:
An Open Source AI Model that outperformed:
- Llama 3.2
- Gpt4o
- o1 mini
Here's everything you need to know about Chinese Deepseek Model:
🧵
An Open Source AI Model that outperformed:
- Llama 3.2
- Gpt4o
- o1 mini
Here's everything you need to know about Chinese Deepseek Model:
🧵
2
DeepSeek is a Chinese-developed open-source LLM series of models.
I would say this is the first real Open Source model.
It has provided details about what data and techniques it used to train the model.
MIT licence, which means you can train and build AI commercial products.
I would say this is the first real Open Source model.
It has provided details about what data and techniques it used to train the model.
MIT licence, which means you can train and build AI commercial products.
3
Deepseek now have 2 flagship models:
1. Deepseek-v3
2. Deepseek r1
Deepseek v3 is is competing with GPT4o and Clade Sonnet 3.5.
v3 outperformed GPT4o and Sonnet 3.5 in MMLU benchmark (to understand human language)
and It is 10x cheaper than GPT$o and Sonnet 3.5.
Deepseek API pricing:
Input: $0.27/million tokens ($0.07/million tokens with cache hits) and output: $1.10/million tokens
Claude 3.5 Sonnet API pricing:
Input costs $3.00 per million tokens, and output costs $15.00 per million tokens.
GPT4o API Pricing:
Input costs $2.50 per million tokens, and output costs $10.00 per million tokens.
Note: Sonnet 3.5 is still the best Coding model and GPT4o is great at logic and Maths.
1. Deepseek-v3
2. Deepseek r1
Deepseek v3 is is competing with GPT4o and Clade Sonnet 3.5.
v3 outperformed GPT4o and Sonnet 3.5 in MMLU benchmark (to understand human language)
and It is 10x cheaper than GPT$o and Sonnet 3.5.
Deepseek API pricing:
Input: $0.27/million tokens ($0.07/million tokens with cache hits) and output: $1.10/million tokens
Claude 3.5 Sonnet API pricing:
Input costs $3.00 per million tokens, and output costs $15.00 per million tokens.
GPT4o API Pricing:
Input costs $2.50 per million tokens, and output costs $10.00 per million tokens.
Note: Sonnet 3.5 is still the best Coding model and GPT4o is great at logic and Maths.
4
Deepseek r1 model: The Reasoning model
Does it beat OpenAI's o1 Model?
I think it sits it between o1-mini and o1 model (better than o1 mini at everything and beats o1 in few things but not all)
But it is 20x cheaper than o1 model so that is massive.
Deepseek r1 pricing:
$0.14 / 1M input tokens (cache hit)
$0.55 / 1M input tokens (cache miss)
$2.19 / 1M output tokens
OpenAI o1 API pricing:
$15.00 / 1M input tokens
$7.50 / 1M cached* input tokens
$60.00 / 1M output tokens
OpenAI o1 Mini API pricing:
$3.00 / 1M input tokens
$1.50 / 1M cached* input tokens
$12.00 / 1M output** tokens
Does it beat OpenAI's o1 Model?
I think it sits it between o1-mini and o1 model (better than o1 mini at everything and beats o1 in few things but not all)
But it is 20x cheaper than o1 model so that is massive.
Deepseek r1 pricing:
$0.14 / 1M input tokens (cache hit)
$0.55 / 1M input tokens (cache miss)
$2.19 / 1M output tokens
OpenAI o1 API pricing:
$15.00 / 1M input tokens
$7.50 / 1M cached* input tokens
$60.00 / 1M output tokens
OpenAI o1 Mini API pricing:
$3.00 / 1M input tokens
$1.50 / 1M cached* input tokens
$12.00 / 1M output** tokens
5
What Deepseek did best is it gave access to its model easily.
You can go to chat .deepseek .com and try the model just like ChatGPT. (Good UI/UX)
You can access their API on patform .deepseek .com and it's comaptable with OpenAI SDK so simple integration.
I will be integrating Deepseek r1 to @CodeGuidedev for documentation reasoning.
You can go to chat .deepseek .com and try the model just like ChatGPT. (Good UI/UX)
You can access their API on patform .deepseek .com and it's comaptable with OpenAI SDK so simple integration.
I will be integrating Deepseek r1 to @CodeGuidedev for documentation reasoning.
6
Day 4/6 of CodeGuide 2.0
✅Day 1: Lauched AI Agent - Codie
✅Day 2: Launched built-in community
✅ Day 3: Mobile/Desktop Shortcut, Notifications,
✅ Day 4: Deepseek r1 integration + Documentation 2.0
Now you'll see implementation-plan .md doc too. This is a step-by-step plan for LLM models to start from step 1 and complete all 30 40 steps to get the app coded.
✅Day 1: Lauched AI Agent - Codie
✅Day 2: Launched built-in community
✅ Day 3: Mobile/Desktop Shortcut, Notifications,
✅ Day 4: Deepseek r1 integration + Documentation 2.0
Now you'll see implementation-plan .md doc too. This is a step-by-step plan for LLM models to start from step 1 and complete all 30 40 steps to get the app coded.
7
Key Takeaways:
Deepseek is a great start for actual Open Source AI models. (Llama lost here)
It is also competing neck to neck with o1, Sonnet 3.5 and GPT4o models (which are closed source)
DeepSeek R1:
o1 performance✅
open source ✅
MIT licence ✅
Future is bright for AI.
Deepseek is a great start for actual Open Source AI models. (Llama lost here)
It is also competing neck to neck with o1, Sonnet 3.5 and GPT4o models (which are closed source)
DeepSeek R1:
o1 performance✅
open source ✅
MIT licence ✅
Future is bright for AI.
8
Join codeguide.dev if you're building with AI tech.
✅ Create Project Development Docs
✅ Use Codie - AI Agent to get coding process help
✅ Join Community of AI Coders
✅ Enjoy Exclusive deals with bolt .new, create AI, softgen AI
More deals and features are coming.
✅ Create Project Development Docs
✅ Use Codie - AI Agent to get coding process help
✅ Join Community of AI Coders
✅ Enjoy Exclusive deals with bolt .new, create AI, softgen AI
More deals and features are coming.