Qi Liu

Cofounder @RekaAILabs, Assistant Professor @HKUniversity Past: @DeepMind, FAIR (@MetaAI), @MSFTResearch, PhD @UniofOxford

[Reka @RekaAILabs]

📢 We are open sourcing ⚡Reka Flash 3.1⚡ and 🗜️Reka Quant🗜️.

Reka Flash 3.1 is a much improved version of Reka Flash 3 that stands out on coding due to significant advances in our RL stack. 👩‍💻👨‍💻

Reka Quant is our state-of-the-art quantization technology. It achieves near-lossless compression of Reka Flash 3.1 to 3.5 bits. 💻

[Reka @RekaAILabs]

We're building the future of intelligence across all modalities. We have raised $110M of new capital from new and existing investors. We are hiring across all roles. 🚀

If you're a creator, developer, or enterprise, explore Reka Vision today! Create reels, search for clips, or ask questions about a video.

Try it yourself: https://t.co/pm7nSgJjjl
Learn more: https://t.co/5R7h1PHabF
Talk to us on Discord: https://t.co/skll3R3SxP
Check out our new website: https://t.co/TxbyPsCe4T

[Reka @RekaAILabs]

Excited to introduce Reka Vision, an agentic visual understanding and search platform. Transform your unstructured multimodal data into insights and actions.

[Reka @RekaAILabs]

⚡We are open sourcing Reka Flash 3, our new reasoning model that was trained from scratch. It achieves competitive performance with only 21B parameters. ⚡

Reka Flash 3 powers Nexus, our new enterprise intelligence platform. 🤖

[Reka @RekaAILabs]

🎉We are excited to introduce Space, a new AI-powered canvas that redefines how you interact with our multimodal models.

Most of our interactions with AI today are carried out in a chat interface. Many of you have experienced our models via Reka Chat.. As our models become more and more capable, we are excited to explore different ways people interact with AI models.

🎯Our goal is to provide a better interface to allow agents to remember past interactions, learn from them, perform actions, and collaborate with multiple humans and AIs. As such, Space will be a gamechanger.

[Guohao Li 🐫 @guohao_li]

@RekaAILabs We integrated Reka's multimodal models which would be helpful for building our multi-agent systems with multimodality!

[Qi Liu @leuchine]

Our new Reka Flash, a 21B multimodal assistant. Try it for free at https://t.co/SRBtHPtApS.

[Reka @RekaAILabs]

We have been busy the past few months and have some exciting updates!📢

We have a new version of Reka Flash⚡️, our powerful 21B model that supports interleaved multimodal inputs (text📄, image🖼️, video📽️, audio🎧).

This update brings significant capability improvements on multimodal understanding, general reasoning, agent building blocks, and speech input-output to make our models more useful across a wide range of practical use cases.

[Shutterstock @Shutterstock]

We’re looking forward to our multi-year partnership with @RekaAILabs! We’re excited to leverage Reka’s AI capabilities to further enrich the metadata and tagging of our content library. Read more: https://t.co/1CtAOQWnAJ

[Arena.ai @arena]

Exciting update - the latest leaderboard result is here!

We have collected fresh 30K votes for three new strong models from @RekaAILabs Reka Core and @Alibaba_Qwen Qwen Max/110B!

Reka Core has climbed to the 7th spot, matching Claude-3 Sonnet and almost Llama-3 70B! Meanwhile, Qwen Max / 110B also achieves top-10/13.

Big congrats Reka and Qwen for the great achievements!

[Reka @RekaAILabs]

Evals are notoriously difficult to get right but necessary to move the field forward. 🌟

As part of our commitment to science, we’re releasing a subset of our internal evals. 🙌

Vibe-Eval is an open and hard benchmark comprising 269 image-text prompts for measuring the progress of multimodal chat models.

Vibe-Eval has two goals:
🔹vibe-checking existing models on day to day tasks 🔹deeply probing into the capabilities of frontier models

These very difficult prompts were constructed by AI experts (us 🤭) with the intention to deliberately find examples where our model (Reka Core) is unable to solve.

On 50% of the prompts in the hard set, none of the existing models (including frontiers) can fully solve. 😲.

We benchmark 13 representative multimodal models using both human evaluation and automatic evaluation.

We also share our findings and insights about the challenges of curating and evaluating hard prompts.

Check out our paper and resources below ⬇️

[Qi Liu @leuchine]

We are thrilled to announce the release of Reka Core. It is available at the Reka Playground https://t.co/SRBtHPtApS. For more details, please refer to our tech report https://t.co/N8wYhQcuEn

[Reka @RekaAILabs]

Meet Reka Core, our best and most capable multimodal language model yet. 🔮

It’s been a busy few months training this model and we are glad to finally ship it! 💪

Core has a lot of capabilities, and one of them is understanding video --- let’s see what Core thinks of the 3 body trailer.👇

[Snowflake @Snowflake]

As a part of our commitment to helping our customers unlock the power of #AI on all types of data, we’re furthering our partnership with @RekaAILabs to bring gen AI to images, video and more to Snowflake Cortex.

Learn more about our partnership:
https://t.co/ZkvLB7Cx1T

[Qi Liu @leuchine]

Honored to be working with our small yet incredibly capable team. Please try it at Reka playground https://t.co/SRBtHPtApS (supports text, images, and videos now).

[Reka @RekaAILabs]

Introducing Reka Flash, our efficient and highly capable multimodal language model.

Try it at Reka playground 🛝 for free today.

🧵 Thread, blog & links below 👇

[Tao Yu @taoyds]

Beyond our Lemur: OPEN LMs for language agents

Introducing 💥OpenAgents💥: an OPEN platform for language agents in the wild!

Analyze data, call plugins, control your browser as ChatGPT Plus, but with OPEN SOURCE code!!

📑: https://t.co/DxPYo1Dq4i
Code: https://t.co/ogTSUbdUVN

[Qi Liu @leuchine]

@Umberto_Senpai Thanks. I think it means "wisdom" in different languages

[Qi Liu @leuchine]

Excited to share the launch of Yasa-1! An incredible work finished by our team of fewer than 20 in just 6 months.

[Reka @RekaAILabs]

We are excited to announce the 1st version of our multimodal assistant, Yasa-1, a language assistant with visual and auditory sensors that can take actions via code execution 🪄.

Yasa-1 can understand text, images, videos, sounds & more! 🚀

Check out more details below👇

[Xterio @XterioGames]

🚨XTERIO AI PLANS REVEALED 🎙️

We're building the future of Web3 x AI
Don't fall behind. Bring a friend 👉🏷️

🔹@DaniYogatama, CEO, Reka
🔹Michael Tong @fat_leopard, MD, Xterio
🔹Edge Wu, CTO, FunPlus
🔹Michael Chen, Managing Director, DST Global

https://t.co/wkmXyoXGHI

[Reka @RekaAILabs]

Check out a video demo of running Reka's multimodal assistant in Snowpark Container Services! 😀

https://t.co/pzIP3J8tFu

[Qi Liu @leuchine]

@matthen2 @RekaAILabs Honored to work wit you in the team!

[Qi Liu @leuchine]

@hai_t_pham @RekaAILabs Great to have you in the team!