Anthropic’s Claude 2 is probably one of the closest competitors that OpenAI’s ChatGPT has. Now, the company has made its AI chatbot even more appealing for users, upping the number of tokens it can translate at one time. The new Claude 2.1 update, now available, ups the total token context window to 200K, allowing the bot to translate roughly 150,000 words of material at once.
On top of adding more context tokens, the upgrade also has halved Claude’s hallucination rate – the rate at which it creates answers and facts – something AI like Claude and ChatGPT have long struggled with.
That 200K context token limit will allow for a lot of room to work in the AI-powered chatbot, too. 200K context tokens are more than enough to cover 500 pages of material, and according to Anthropic, the Claude 2.1 upgrade is now able to handle uploading and analyzing the entirety of Homer’s The Odyssey.
This is a huge improvement, as the context that an AI bot like Claude and ChatGPT can pick up on greatly determines just how useful the entire system is, especially for longer-form projects. Anthropic says that uploading and analyzing that amount of content still takes several minutes. However, the company expects the processing time to improve as the technology does.
Additionally, Anthropic revealed that the recent Claude 2.1 update now sees the chatbot making 30 percent fewer errors in extremely long documents, and that it has a three to four times lower rate of “mistakenly concluding a document supports a particular claim” than it used to, even when using those larger and more robust context windows.
This upgrade will no doubt continue to make Claude 2 an ideal alternative to ChatGPT, and if you’ve been looking to take advantage of the power of AI chatbots, but haven’t picked up ChatGPT Plus just yet, then Claude 2.1 should be a solid option, at least until OpenAI improves ChatGPT and re-opens ChatGPT Plus subscriptions.
