Anthropic May Soon Launch Claude 3.5 Haiku to Regular Users

Recently, a user on X spotted Anthropic’s Haiku 3.5 on the Claude.ai application, which may hint towards a public release in the coming few days. The Haiku 3.5 is the latest version, of the smallest variant in Anthropic’s family of AI models. 

Tibor Blaho, known for spotting unreleased features in the back end, was at it again. It looks like they were trying to upload an image using Haiku, but the warning message read that ‘Claude 3.5 Haiku does not support images’. Further, images with the backend code were also shared, which revealed the model’s code“claude-3-5-haiku-20241022”.

New Claude experiments: "claude_ai_chili" and "claude_ai_cinnamon" - claude-3-5-haiku-20241022 is coming to claude[.]ai (no vision capabilities & no PDF uploads) - and "hidden" as "Claude 3 Haiku"? pic.twitter.com/ujVkjbFQwr

— Tibor Blaho (@btibor91) December 6, 2024

Last month, Anthropic made the Claude 3.5 Haiku model available for API-only use on AWS Bedrock and Google’s Vertex AI. Anthropic also mentioned that 3.5 Haiku is a text-only model, with image inputs to follow soon. That said, not only is it significantly better than Haiku 3, but Claude’s 3.5 Haiku outperforms the former flagship level Opus 3 model and is also the fastest model to date. 

Claude 3.5 Haiku is better than the older Claude 3.5 Sonnet and Claude 3 Opus.
I don’t know how Anthropic does it, but they always have the best model for coding. Incredible work, especially considering it’s more affordable, this is huge for developers pic.twitter.com/7hXDOgJXAP

— AshutoshShrivastava (@ai_for_success) November 4, 2024

If Haiku 3.5 is available to users soon, it will be great news. Developers have already been able to test its capabilities, and Sean Ward, CEO of iGent AI said "Claude 3.5 Haiku shows impressive reasoning and code generation abilities, including demonstrating powerful multi-turn code refinement which reduced code-related errors by 60%, and puts it in the same territory as substantially bigger models”

Recently, Anthropic announced a new feature that integrates Google Docs with Claude to ‘directly reference’ content inside your chats. You can insert a Google Doc file in a chat or a project, which allows Claude to access the same and use the information to generate outputs. Moreover, you can also set Claude’s output style and add pre-defined instructions and preferences for the output. 

Dario Amodei, in a podcast episode with Lex Fridman also confirmed that there are active plans to release the flagship Opus 3.5 model soon. Anthropic recently announced another round of funds, raising $4 billion from Amazon Web Services (AWS), bringing Amazon’s investments in the AI startup to $8 billion.