While there are continuous developments in the field of artificial intelligence, Anthropic has taken an important step towards making its artificial intelligence models more capable. Accordingly, the company has updated its models in the Claude family, especially the Claude 3.5 Sonnet.
What does the renewed Claude 3.5 Sonnet offer?
Anthropic aims to offer a more capable artificial intelligence by sharing important updates in its Claude models. The company’s most powerful Claude 3.5 Sonnet model achieved its highest score to date by achieving a score of 49.0 percent in the SWE-bench Verified benchmark. This is a significant increase from the 33.4 percent score of the previous Claude 3.5 Sonnet model.
Double-digit percentage point increases were also recorded in other benchmarks such as TAU-bench and GPQA. Anthropic also stated that the updated model outperformed rival models such as the Google Gemini 1.5 Pro in some evaluations.
Anthropic also announced that it has made the smaller and more affordable Claude 3.5 Haiku model available. Although it was initially only text-based, it managed to catch up with the largest model of the previous generation, Claude 3 Opus. This makes it an alternative to models such as GPT-3 Mini and Gemini Flash.
The renewed 3.5 Sonnet model has been tested in collaboration with organizations such as the US and UK Artificial Intelligence Security Institutes (AISI). Anthropic also stated that the model continues to comply with various scaling frameworks and ASL-2 security standards.
Both Claude models will continue to be offered at the same prices through the Anthropic API and cloud integrations.
{{user}} {{datetime}}
{{text}}