Anthropic announced today that the 1M token context window for Claude Opus 4.6 and Sonnet 4.6 is now generally available, with the long-context pricing premium completely eliminated.
The 1M token window - roughly 750,000 words, or the equivalent of about ten full-length novels - has been available in beta since both models launched in February 2026. But until now, API requests exceeding 200K tokens carried a surcharge: roughly 2x on input tokens and 1.5x on output. A long-context Opus request that would normally cost $5 per million input tokens jumped to $10.
That surcharge is gone. A 900K-token request now costs exactly the same per-token rate as a 9K-token request.
600 Images Per Request and Sharper Retrieval
The GA release also bumps the maximum number of images and PDF pages per request from 100 to 600. For anyone building document processing workflows or analyzing large codebases, that's a practical capacity jump.
On the performance side, Anthropic says Opus 4.6 scores 76% on the 8-needle MRCR v2 benchmark - a test measuring how well models retrieve specific information buried deep in very long contexts - compared to 18.5% for the previous Sonnet 4.5. In plain terms: Opus 4.6 is substantially better at actually using all that context, not just accepting it.
Who Benefits Most
The pricing change matters most for API developers building applications that process long documents, large codebases, or extended conversation histories. The previous 2x surcharge made 1M-context applications expensive to run in production. At standard rates, those same applications just got a significant cost cut.
For claude.ai subscribers, Sonnet 4.6 with 1M context is available on the free tier. Opus 4.6 requires Pro ($20/month) or Max ($100-200/month). Claude Code users on Max, Team, or Enterprise plans also get the full 1M window.
The move puts direct pricing pressure on OpenAI's GPT-4.1, which offers a 1M context window at $2/$8 per million tokens. Anthropic's Sonnet 4.6 at $3/$15 is in the same ballpark, while Opus 4.6 at $5/$25 is positioned as the premium option with stronger long-context accuracy. For developers already committed to Claude's API, this is a straightforward win. For those shopping between providers, the decision just got more interesting.