Reasoning Benchmark

ForecastBench

ForecastBench: LLM forecasting benchmark scored with a Brier score where LOWER is better (0 = perfect calibration, larger = worse). Scores here are Brier scores as printed by Cloudflare's internal Decision Index 0.2.1 run (e.g. Clef 13.9, Clef-flash 10.6).

Category

Reasoning

Score type

brier_lower_is_better

Scale

100.0

lower is better

Models tested

2

What is ForecastBench?

ForecastBench is a benchmark for general reasoning and comprehension. The official description says: "ForecastBench: LLM forecasting benchmark scored with a Brier score where LOWER is better (0 = perfect calibration, larger = worse)". Below we describe in simple words what benchmarks of this type generally examine.

What does it test?

ForecastBench measures general reasoning and comprehension. In detail, it checks:

  • drawing the right conclusion from a passage or scenario
  • commonsense judgement about everyday situations and words
  • counting, comparing and reasoning over several facts at once
  • avoiding shortcuts that memorise answers instead of reasoning

How to read the score

The score is a Brier score for probabilistic forecasts: it punishes overconfident wrong predictions. Unlike the other score types, lower is better, and 0 is a perfect forecast.

Official methodology note: ForecastBench: LLM forecasting benchmark scored with a Brier score where LOWER is better (0 = perfect calibration, larger = worse). Scores here are Brier scores as printed by Cloudflare's internal Decision Index 0.2.1 run (e.g. Clef 13.9, Clef-flash 10.6).

Frequently asked questions

What is the ForecastBench benchmark?

ForecastBench is a benchmark for general reasoning and comprehension. The official description says: 'ForecastBench: LLM forecasting benchmark scored with a Brier score where LOWER is better (0 = perfect calibration, larger = worse)'. Below we describe in simple words what benchmarks of this type generally examine.

What does ForecastBench test?

ForecastBench mainly measures general reasoning and comprehension. Concretely, it checks: drawing the right conclusion from a passage or scenario; commonsense judgement about everyday situations and words; counting, comparing and reasoning over several facts at once.

How is the ForecastBench score calculated?

The score is a Brier score for probabilistic forecasts: it punishes overconfident wrong predictions. Unlike the other score types, lower is better, and 0 is a perfect forecast. The declared maximum score of this benchmark is 100.0.

Where do I get ForecastBench data and scores?

The official source is https://clef-evals.workers-ai-mle.workers.dev/. The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/forecastbench/scores) or pull it directly from the Open LLM Radar API with any HTTP client.

Model scores 2 models tested

Reset
# Model Score Date Source Verified
1 Clef-Flash
Cloudflare
10.6 Brier
— Report Verified
2 Clef
Cloudflare
13.9 Brier
— Report Verified

What now?

You've seen how models score on ForecastBench — here's what to do next.