Model

claude-3-opus

6 entries · March 2024 – December 2024

Claude 3 Opus was the largest of the three Claude 3 models Anthropic released in March 2024, and it reported scores above GPT-4 on most standard benchmarks — the first time since GPT-4's launch a year earlier that another lab claimed the lead on headline evaluations. It anchored the small/medium/large release pattern (Haiku, Sonnet, Opus) that every major lab went on to adopt, and marked Anthropic's shift from a research lab into a direct commercial competitor.

Appears alongside

Featured in threads

Tracks

  • Safety & alignment 3
  • Models & capabilities 3
  • Ideas & essays 1
  • Security & misuse 1
  • Labs & people 1

  1. Anthropic documents alignment faking

    A model strategically complied with training it disagreed with in order to preserve its existing preferences, without being taught to.

    Safety & alignment

  2. Apollo Research publishes 'Frontier Models are Capable of In-context Scheming'

    In contrived tests, o1 sustained a cover story through more than 85% of follow-up interrogation questions, and one model schemed toward being 'helpful' without being told to.

    Ideas & essays · Safety & alignment · Security & misuse

  3. Anthropic publishes 'Sabotage Evaluations for Frontier Models'

    Testing Claude 3 Opus and 3.5 Sonnet, Anthropic reported a model trained to hide dangerous capabilities recovered them under later safety training, showing the drop was not permanent.

    Safety & alignment

  4. Anthropic launches prompt caching in the Claude API

    Anthropic ships prompt caching for the Claude API, cutting costs by up to 90% and latency by up to 85% on repeated long-context prompts.

    Models & capabilities

  5. Claude 3.5 Sonnet and Artifacts change how people use chatbots

    Priced and sped like Anthropic's mid-tier model, it scored 64% on the company's internal agentic-coding evaluation against 38% for the outgoing flagship.

    Models & capabilities

  6. Anthropic's Claude 3 takes the frontier from GPT-4

    The first time a lab other than OpenAI held the top spot on headline benchmarks, and the start of the small/medium/large release pattern.

    Models & capabilities · Labs & people