Anthropic AI R&D - Time Series Forecasting MSE: leaderboard
Time-series forecasting task from Anthropic's internal AI R&D rule-out suite: mean squared error on the hard variant, where the card equates below 5.3 with 40 hours of human effort; lower is better. From Table 2.3.5.A of the Claude Opus 5 System Card (July 2026); the Sonnet 5 row is from Table 2.3.3.A of the Sonnet 5 card.
Metric: Mean Squared Error (Hard Variant). Source: www-cdn.anthropic.com. Saturation forecast: Around February 2028. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Mythos 5 | 4.51 | #8 |
| 2 | Claude Opus 4.7 | 4.78 | #51 |
| 3 | Claude Opus 5 | 5.68 | #16 |
| 4 | Claude Sonnet 5 | 5.8 | #65 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=anthropic-ai-r-d-time-series-forecasting-mse · How It Works · Data refreshed daily, snapshot 2026-11-05.