Jofie Yordan · · 1 min readSingaporean ex-Googler’s AI startup launches new model to rival Bard, ChatGPT
Photo credit: Shutterstock
Reka, a California-based AI startup, appears to be ready to go head-to-head with the likes of Google and Microsoft. It has rolled out its latest multimodal and multilingual language model, Reka Flash.
The company was founded in 2022 by former researchers from Google DeepMind and Meta, including Yi Tay, a Singaporean who serves as Reka’s chief scientist. The new launch follows a US$58 million funding round in June 2023.
According to the firm, Reka Flash has 21 billion parameters. It said that the model is competitive with Google Bard’s Gemini Pro and OpenAI’s GPT-3.5 across different benchmarks.
Reka Flash has been trained on text from over 32 languages. According to Reka’s evaluations, it outperforms larger models like Llama 2, Grok-1, and GPT-3.5 in aspects such as reasoning, code generation, and question answering.
Additionally, the company released a more compact version of the model called Reka Edge. This variant, which has 7 billion parameters, is tailored for scenarios with limited resources, making it ideal for on-device or local applications.
Both Reka Flash and Reka Edge are currently available for beta testing. The company also said it is set to launch its largest model, Reka Core, in the coming weeks.
See also: The good, the bad, and the hallucinations: What SEA startups think of AI
Editing by Miguel Cordon and Lorenzo Kyle Subido
(And yes, we’re serious about ethics and transparency. More information here.)
Stay updated on the go with our mobile app.
Get latest insights with smoother, more personalized experience through TIA mobile app.
