Anthropic Unveils Upgraded Claude 3.5 Models

www.analyticsdrift.com Image source: Analytics Drift

Introducing Two New Claude 3.5 Models

[{"selector":"#anim-2d646f58-6ec7-4e10-a698-44a7bc7b36d6","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-a49d8db3-fa7b-4637-92cc-508a81bd780c","keyframes":{"transform":["translate3d(0px, 176.35067%, 0)","translate3d(0px, 0px, 0)"]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-ec14b53a-0e0c-4c2f-af73-819f7fb3fcde","keyframes":{"opacity":[0,1]},"delay":120,"duration":1300,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-40c0e1ae-543a-4d56-9a6d-0ad0d0bcbbca","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] On Tuesday, 22nd October 2024, Anthropic announced the release of two new Claude 3.5 models: Claude 3.5 Sonnet (Upgraded) and Claude 3.5 Haiku. Image source: Claude

Enhancement in the Performance of Predecessors

[{"selector":"#anim-c98096d5-0480-48b9-aaa1-a5d6fe41b476","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-c019d9e3-95a2-4285-bb9b-911a2ac9ca70","keyframes":{"transform":["translate3d(0px, 163.52842%, 0)","translate3d(0px, 0px, 0)"]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-1059a786-3f43-44aa-88dd-2b25b94fd683","keyframes":{"opacity":[0,1]},"delay":120,"duration":1300,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-b77d71ae-2c17-48f2-b8ab-d906f1ff8436","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] Both the models, Claude 3.5 Sonnet and Haiku, significantly improve the performance of the previously deployed models, with 3.5 Sonnet offering computer use functionality. Image source: Canva

Computer Use Feature to Automate Tasks

[{"selector":"#anim-93d9b414-9229-4486-a665-600e1d6c8591","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-0abc93bb-b929-4ce3-b587-d94113cce4d4","keyframes":{"transform":["translate3d(0px, 154.75146%, 0)","translate3d(0px, 0px, 0)"]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-ec26042a-9169-4433-82c8-f2dfd28f9709","keyframes":{"opacity":[0,1]},"delay":120,"duration":1300,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-4663e175-9736-4c0c-9581-dea45059d55d","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] The computer use feature offered by Claude 3.5 Sonnet enables you to automate most online processes, such as filing online forms. Using this feature, you can safely allow Claude to interact with your machine to perform various tasks. Image source: Claude

Improved Results for Popular Benchmarks

[{"selector":"#anim-29a3413e-6d4d-481c-ba56-38ec2a9c7162","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-f1db02a5-2c60-4d1e-8eaf-99cb94c7654f","keyframes":{"transform":["translate3d(0px, 153.43565%, 0)","translate3d(0px, 0px, 0)"]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-122d3020-dd89-44f5-a605-c143cf967bde","keyframes":{"opacity":[0,1]},"delay":120,"duration":1300,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-9f4a98d6-f9dd-4547-9e32-8fcb2c9d5359","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] The popular testing benchmarks highlight a significant impact on the model performance of Claude’s new models. For the SWE-bench Verified tests, 3.5 Sonnet resulted in a staggering increase from 33.4 to 49%. Image source: Claude

Expansion in the Number of Use Cases of Newer Models

[{"selector":"#anim-b4b839bf-225e-48f3-87b8-3cdd9f60112b","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-094d4209-21d4-4ec4-999f-8d63492f8855","keyframes":{"transform":["translate3d(0px, 155.89485%, 0)","translate3d(0px, 0px, 0)"]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-6339db28-60b6-4cf4-8c97-35ae72febcba","keyframes":{"opacity":[0,1]},"delay":120,"duration":1300,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-4561c707-2595-4028-a0fb-2a5047fafa4d","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] Although in the testing phase, companies like Asana, Cognition, Canva, DoorDash, The Browser Company, and Replit have already started working with Claude 3.5 Sonnet. Image source: AD

Offering Cost-Effective AI with Peak Performance

[{"selector":"#anim-0acaadf9-1b8f-4f23-b6e7-ea8e52d43d92","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-e420ec0c-0ec4-48cd-88f1-df879802c605","keyframes":{"transform":["translate3d(0px, 180.93782%, 0)","translate3d(0px, 0px, 0)"]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-c5323e07-6b5d-4d68-9634-1c21cb5763cb","keyframes":{"opacity":[0,1]},"delay":120,"duration":1300,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-beddcd27-5d43-4b3a-9b11-1e00789e498e","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] The Claude 3.5 Haiku surpasses the previously used Claude 3 Opus model in numerous benchmarks, providing improved performance and speed at a reasonable price. Image source: Canva

Revolutionizing Automation of Redundant Tasks

[{"selector":"#anim-6cee6f1a-f02e-4951-bb2e-8ccba984faab","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-efce588d-f516-441d-b376-65f6271c05c3","keyframes":{"transform":["translate3d(0px, 176.35067%, 0)","translate3d(0px, 0px, 0)"]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-22520d0d-e621-4471-b4aa-1ef64f07e394","keyframes":{"opacity":[0,1]},"delay":120,"duration":1300,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-49a0a7cc-ba47-4710-95b5-a3d5abfcebbe","keyframes":{"opacity":[0,1]},"delay":120,"duration":1200,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] The new Anthropic releases aim to automate most daily tasks at an affordable price, marking a new leap in general artificial intelligence. Image source: Canva Read more