Anthropic released Claude Opus 4.7 with a 1M-token context window in the flagship tier, matching Gemini’s spec and tripling the previous Opus ceiling. Pricing on long-context calls stays in the same band as 200K, which is the part operators will react to: the marginal cost curve for agentic workflows that previously hit context ceilings just shifted down.
The model card highlights gains on long-horizon tool-use evals and code-repo retrieval at the upper end of context. Sub-100K performance is reported as flat versus 4.6 , useful detail, because the practical question is whether shops swap their default model or only their long-document pipeline.
Worth tracking: how the API price-per-effective-token compares to running multiple shorter calls with retrieval. Long-context isn’t free, and the right answer depends on the workload shape.