Claude 3.5 Sonnet Solves 64% More Coding Problems Than Opus 3, Highlighting Rapid LLM Evolution

Anthropic's latest advancements in large language models (LLMs) have demonstrated significant performance gains, with the recently released Claude 3.5 Sonnet model outperforming its higher-tier predecessor, Claude 3 Opus, in crucial benchmarks. This rapid iteration challenges the notion that LLM progress is always linear, as observed in a recent social media post.

Released in June 2024, Claude 3.5 Sonnet, a mid-tier model, has set new industry standards, particularly in coding proficiency. In an internal agentic coding evaluation, Claude 3.5 Sonnet successfully solved 64% of problems, a substantial improvement over Claude 3 Opus, which solved 38%. Anthropic also stated that Claude 3.5 Sonnet surpassed Claude 3 Opus on standard vision benchmarks, especially in tasks requiring visual reasoning like chart interpretation.

Anthropic, backed by tech giants like Amazon and Google, consistently aims to enhance the trade-off curve between intelligence, speed, and cost with frequent updates. The company's strategy involves releasing new models that often supersede previous versions, even across different tiers, to deliver more capable and efficient AI. This iterative approach is a hallmark of the competitive LLM development landscape.

However, this rapid evolution can lead to perceptions of older models being "retired" or even "worse" in certain aspects, as noted by a user on social media: > "Anthropic retiring Opus 3, or Sonnet 3.5, does kinda seem to mean LLMs have just gotten worse at some hard-to-define X that Opus or Sonnet were good at," stated 1a3orn in a tweet. While Anthropic has not officially announced a functional "retirement" of these models in a way that implies degradation, the introduction of superior successors effectively shifts user adoption.

The company has continued its aggressive development roadmap, with plans to release Claude 3.5 Haiku and an updated Claude 3.5 Opus later in 2024. Furthermore, Claude 4 models, including Opus 4 and Sonnet 4, were introduced in May 2025, demonstrating Anthropic's commitment to continuous innovation in the AI space. These developments underscore a dynamic environment where advancements are constant, and the capabilities of LLMs are rapidly redefined.