Yi-01.AI

A global company building generative AI LLM and applications

[Kai-Fu Lee @kaifulee]

Starting your week with more slides to write? Our @popai team just dropped "Slide Agent" to make presentations simple with AI 👉 Prompt > select template (300+) > AI generated draft > reformat (layout, charts, images, logos) > download in .pptx & edit to perfection. Most presentation tools or agents either stop at generating drafts or only do well in formatting. Consider PopAi a "ChatGPT+Canva" in one.

PopAi https://t.co/41s4aGHC5o
Slide Agent full demo https://t.co/DE1yLzxgeR

[Kai-Fu Lee @kaifulee]

The biggest revelation from Deepseek is that Open Source has won. For a 1% difference in performance, it will be difficult for OpenAI to justify its price when the competition is free and formidable. -from my interview with Bloomberg

[Kai-Fu Lee @kaifulee]

DeepSeek is becoming a Windows kernel demanded by businesses, but https://t.co/5Nq9vwwj3V is aspired to build the Windows system and interface to ignite it. Check out more on: https://t.co/Orm87AbQuz

Thanks @BloombergTV @DavidInglesTV and @BelleDroulers for the insightful interview.

[Rohan Paul @rohanpaul_ai]

Chinese startup 01 .ai trains competitive LLM using 95% fewer resources through innovative engineering optimization.

01 .ai trained a GPT-4 competitor using just 2,000 GPUs and $3M, while achieving competitive performance. Through innovative engineering and optimization techniques, they achieved what OpenAI did with $80-100M, demonstrating remarkable cost efficiency in LLM training.

→ Training Resource Optimization at 01 .ai

Using only 2,000 GPUs versus OpenAI's estimated 10,000+ GPUs for GPT-3. The company achieved competitive performance despite severe hardware constraints due to US regulations.

→ Cost Efficiency Breakthrough

$3M total training cost compared to OpenAI's $80-100M for GPT-4. Model ranked sixth in performance according to UC Berkeley's LMSIS benchmark.

→ Technical Innovation in Inference

Transformed computational problems into memory-oriented tasks. Built multi-layer caching system and specialized inference engine. Achieved inference costs of 10 cents per million tokens - 1/30th of industry standard.

→ Engineering Focus Areas

Prioritized GPU resource allocation. Optimized both training speed and inference efficiency. Developed custom inference architecture for maximum hardware utilization.

[Yi-01.AI @01AI_Yi]

Yi and Yi 1.5 are evolving🌳

[Omar Sanseviero @osanseviero]

The (non-exhaustive) evolution of base models

If you want to learn more about it and how to use these models, check out the freshly released book "Hands-On Generative AI", written with @pcuenq @multimodalart and @johnowhitaker!

https://t.co/tx9vnyHGzC

[Kai-Fu Lee @kaifulee]

https://t.co/tq9WCFxkLe trained the #6 model in the world for $3M pre-train cost. And the inference price is $0.14/million tokens! https://t.co/a60oSfIBeR

[Yi-01.AI @01AI_Yi]

🎉Love seeing Yi models in @CamelAIOrg! Powerful Yi models join forces with this awesome multi-agent framework. Can't wait to see what AI agents you'll build!
#YiLightning #LLM #AI

[CAMEL-AI.org @CamelAIOrg]

📢 We've just added support for the Yi-series of LLM models in the 🐫 CAMEL framework!

This enhancement allows users to leverage various performance tiers with models like yi-lightning, yi-large, yi-medium, and yi-large-turbo, providing greater flexibility in language processing tasks.

Thanks to our contributor MuggleJinx for this significant contribution! 🤝 Explore more here: https://t.co/3RIhavM97P.

[Yi-01.AI @01AI_Yi]

🌍Exciting news from our developer community!

We're thrilled to share a blog on Refactor Earth, which explores an innovative approach to sustainable AI.

By combining Yi-Large and CodeBERT, this project optimizes code for efficiency, achieving over a 10% reduction in its environmental footprint!

🏆 Proud to announce that this project won 1st Place at the GenLab x AI Engineer World's Fair Hackathon. Huge kudos to @Shalini_Ananda for this remarkable achievement!

https://t.co/Jy6yGYYygN

#YiLarge #AI #Hackathon

[Yi-01.AI @01AI_Yi]

Thrilled to see such widespread adoption of Yi!

Huge thanks to @huggingface, @ollama, and mradermacher for your incredible support!

ollama run hf(.)co/mradermacher/Yi-1.5-34B-Chat-16K-GGUF

#Yi34B #LLM #AI

[Julien Chaumond @julien_c]

The @ollama - @huggingface integration has been rolled out for 1 week now, how it’s going?

Obviously, pretty well! We’re having on average 4500 pulls per day. That’s about one pull every 20 seconds!

What’s the top models you may ask?

Llama-3.2-1B still on top thanks to its small size but yet still providing very helpful responses. Try it yourself!

ollama run hf(.)co/bartowski/Llama-3.2-1B-Instruct-GGUF

via @ngxson 🐐 and cc @bartowski1182

[Yi-01.AI @01AI_Yi]

We are proud to present the latest model ⚡️Yi-Lightning ⚡️ now #6 in the world, higher than the original GPT-4o released 5 months ago. Also humbled that @01AI_Yi is ranked #3 LLM player on @lmarena_ai Chatbot Arena -- open after OpenAI, Google and tie with xAI to serve other broader parts of the world under our vision "Make AGI Accessible and Beneficial to Everyone" 💪

[Arena.ai @arena]

Big News from Chatbot Arena!

@01AI_YI's latest model Yi-Lightning has been extensively tested in Arena, collecting over 13K community votes!

Yi-Lightning has climbed to #6 in the Overall rankings (#9 in Style Control), matching top models like Grok-2. It delivers robust performance in technical areas like Math, Hard Prompts, and Coding.

Huge congrats to @01AI_YI!

Meanwhile, GLM-4-Plus by Zhipu AI (@ChatGLM) has also entered the top 10, marking a strong surge for Chinese LLMs. They're quickly becoming highly competitive. Stay tuned for more!

More analysis below👇

[Yi-01.AI @01AI_Yi]

We're thrilled to unveil Yi-Lightning and Yi-Lightning-Lite, our latest proprietary models!

Both are now accessible via API at https://t.co/lQZE6F4mXe and featured in @lmarena_ai's Chatbot Arena (https://t.co/T3IeZXG9gB).

Welcome to give it a try!

[Yi-01.AI @01AI_Yi]

Great to see how easy it is to build a search page with Yi-Coder and Cursor! Check out this useful tutorial!

[Second State @secondstateinc]

Write a Search Webpage with local hosted Yi-coder and Cursor without any coding experience. https://t.co/33kLIgGw3c
✅ Set up Yi-Coder locally
✅ Add it on Cursor
✅ Build a functional search page from scratch using AI
Check out the demo inside!

[Yi-01.AI @01AI_Yi]

Huge thanks to @josephpollack for bringing this amazing demo! Now Yi-Coder's power is at your fingertips!

[Joseph Pollack #Ï 🎗️ @josephpollack]

🙋🏻‍♂️hey there folks,

just released a coding model under 10B parameters with 125K context window , achieving (very high!) SOTA scores on evals. you can try it out for yourself on @huggingface 👇🏻📷
models : https://t.co/9I9qigqTTt

@Gradio demo : https://t.co/F0qUud4sbb

[Yi-01.AI @01AI_Yi]

We hear awesome feedback on our Sep 4 Yi-Coder release and so glad the community finds it helpful! Here's more scoop🍦on our tech blog -- "Meet Yi-Coder: A Small but Mighty LLM for Code" https://t.co/C5Ke4KY7up

[VentureBeat @VentureBeat]

Yi-Coder: The open-source AI that wants to be your coding buddy https://t.co/HIHHYuahHl

[Yi-01.AI @01AI_Yi]

Credit: This demo idea is inspired by LlamaCoder https://t.co/w7cDAJZBb3

[Yi-01.AI @01AI_Yi]

🚀 Yi-Coder is open-sourced!

The 'Small but Mighty' LLM offers SOTA coding performance under 10B parameters. Excel in code editing, completion, debugging, and math reasoning.

✅ 2 sizes: 9B & 1.5B (Chat & Base)
✅ 128K context length
✅ Support 52 programming languages

Explore it now👇
https://t.co/YYA0nReMn4

#YiCoder #01AI #LLMs #OpenSource

[Yi-01.AI @01AI_Yi]

@xlr8harder Yes, apologies, this was a typo on our end!

[Yi-01.AI @01AI_Yi]

🔥 Meet Yi-Large Turbo: the powerful, cost-effective upgrade to Yi-Large. Faster and more affordable at only $0.19 per 1M tokens for input and output. Ideal for heavy data tasks like complex inference and high-quality text generation. Check it out now: https://t.co/TWr3a3s3rE
#AIInnovation #DeepLearning #MachineLearning #LLM