Traictory tracks 22 models from Anthropic. The earliest entry is Claude 3 Opus from February 2024; the most recent is Claude Fable 5.1, released in September 2026. 12 of them shipped in the last twelve months.
The strongest Anthropic model in our data is Claude Opus 4.5, with an average benchmark score of 87.8% across 6 published results, including 81.0% on SWE-Bench and 87.0% on GPQA. Across the 22 Anthropic models with benchmark data the average is 71.7%.
22 of the models are available through an API, with input pricing from $0.25 to $25 per million tokens; Claude 3 Haiku is the cheapest. The largest context window is 1M tokens on Claude Fable 5.1. 21 of the 22 models accept images or other non-text input. None of the listed models publish open weights.
Anthropic Models
Anthropic Model Comparisons
Other Companies
.svg?dpl=dpl_H7DLTAgfxbfxrpouWPhFKgpPk6mq){kind=icon}
Anthropic model catalog updated in 2026. See model comparisons and best models for coding.