Long Context Windows Enabled Adoption

Anthropic won early enterprise adoption by solving the most obvious failure mode in business LLM use, not enough room to include the actual work artifact. In May 2023, Claude jumped to a 100K token context window, while GPT 4 initially exposed 8K in the API, which meant a company could paste in a long contract, policy manual, research report, or codebase slice and ask one model to read the whole thing at once instead of stitching together many smaller prompts.1 2 3 4

• This mattered because early enterprise AI was mostly document work. Legal teams, analysts, support ops, and engineers wanted answers over PDFs, transcripts, and internal docs. A larger window made retrieval and prompting much simpler, because more of the source material could sit directly inside one request.2 4 5
• The wedge also fit Anthropic’s distribution. Anthropic was partnering into enterprise channels with Scale in April 2023 and BCG in September 2023, reinforcing Claude as the model for internal company knowledge work rather than a mass market chat product.3 6 7
• That positioning created a bridge to Anthropic’s later coding surge. Long context is useful for code for the same reason it is useful for documents, because developers need the model to see more files, dependencies, and instructions at once. Claude 3.5 Sonnet later turned that enterprise foothold into faster API growth.5 8

Going forward, long context becomes less of a one time feature and more of a platform expectation. Anthropic’s advantage is that it used context length to get inside high value business workflows first, and once teams build around a model for contracts, research, and code, that usage can expand into broader agentic and developer products.5 8