Opus 4.7, Now Live in :Harvey:

Today, we are making Claude Opus 4.7 available in Harvey. Opus 4.7 is Anthropic's latest frontier model, building on the agentic strengths of Opus 4.6 while delivering improved reasoning calibration, stronger performance on ambiguous tasks, and gains in substantive accuracy across complex legal workflows.

Early access evaluations point to a model that is better at matching its reasoning depth to the complexity of the task at hand. On our BigLaw Bench evaluation suite, Opus 4.7 scored 90.9%, the highest score of any Claude model in Harvey to date — 45% of tasks received perfect scores, with 88% scoring at or above 0.80. We expect Opus 4.7 to excel in deal management, risk assessment, drafting, and document-heavy transactional work where precision and analytical depth are essential.

Our teams highlighted several additional strengths. The model demonstrates notably improved reasoning calibration. For straightforward questions like identifying governing law, it returns direct answers rather than multi-paragraph explanations. On harder analytical tasks, such as distinguishing between assignment provisions and change-of-control provisions, the model performed correctly where prior models have historically struggled.

While substantive accuracy and analytical capability is strong, qualitative evaluators noted some polish gaps relative to Opus 4.6, such as tone mismatch, a tendency to overshoot on detail, and organizational missteps such as non-standard citations.

We'll be rolling out Opus 4.7 in the model selector to eligible clients in the US and EU in the coming days. AU availability will be announced shortly.