What are the multimodal capabilities of Reka Core?
Reka Core excels in both language and vision tasks, performing competitively on benchmarks such as MMLU, GSM8K, and MMMU. It is capable of processing text and images, making it suitable for applications that require image and video understanding.

How does Reka Core compare to models like GPT-4 and Claude 3 Opus?
Reka Core approaches the capabilities of GPT-4 and Claude 3 Opus, demonstrating strong performance in multilingual and long-context tasks. While it may not always surpass these top models, it offers competitive performance and is a viable alternative for developers.

What is the maximum context length for Reka Core?
The maximum context length for Reka Core is 128,000 tokens.

How does Reka Core perform in multilingual tasks?
Reka Core performs exceptionally well in multilingual tasks, often outperforming larger models in this domain, which makes it a strong choice for applications requiring robust language support across various languages.