Scale AI, 2025
PRBench Finance
High-stakes professional reasoning in finance conversations, graded against expert-written rubrics. Finance is one half of PRBench; the other half is Legal.
About this benchmark
- Size
- 600 Finance tasks (plus a 300-task Finance-Hard subset). Across Finance and Legal: 1,100 tasks.
- Public split
- All tasks and rubrics on Hugging Face (not gated)
- Grading
- Weighted rubric score clipped to 0-1, with 10-30 expert criteria per task. The board shows it on a 0-100 scale.
- Licence
- CC-BY-4.0 (data)
- Submission
- Not stated
- Products allowed
- Not stated
- Latest version
- HF last modified 2026-06-24
- What it does not cover
- No tool use, live data retrieval or numeric spreadsheet work.
Results
10 modelsEach score shows who reported it: the benchmark's authors (quoted from their paper), the benchmark's official board, or Moneda (run with the benchmark's own grader and rules). We do not copy other aggregators.
PRBench Finance HF last modified 2026-06-24
weighted rubric score (0-100), higher is better. 10 models.
Updated date not stated
- Muse Spark 1.359.5
- Muse Wonder 1.058.7
- Muse Spark 1.155.0
- Claude Fable 553.9
- claude-opus-4-6 (Non-Thinking)53.3
- Muse Spark52.4
- gpt-551.3
- gpt-5-pro51.1
- Fable 5.150.8
- gpt-5.6-sol (max)50.5
PRBench Finance HF last modified 2026-06-24
weighted rubric score (0-100), higher is better. 10 models.
Updated date not stated
- 1Muse Spark 1.359.5
- 2Muse Wonder 1.058.7
- 3Muse Spark 1.155.0
- 4Claude Fable 553.9
- 5claude-opus-4-6 (Non-Thinking)53.3
- 6Muse Spark52.4
- 7gpt-551.3
- 8gpt-5-pro51.1
- 9Fable 5.150.8
- 10gpt-5.6-sol (max)50.5
| 1 | Muse Spark 1.3 | Meta | 59.5 | Default | Scale Labs leaderboard Official board, date not stated |
|---|---|---|---|---|---|
| 2 | Muse Wonder 1.0 | Meta | 58.7 | Default | Scale Labs leaderboard Official board, date not stated |
| 3 | Muse Spark 1.1 | Meta | 55.0 | Default | Scale Labs leaderboard Official board, date not stated |
| 4 | Claude Fable 5 | Anthropic | 53.9 | Default | Scale Labs leaderboard Official board, date not stated |
| 5 | claude-opus-4-6 (Non-Thinking) | Anthropic | 53.3 | Default | Scale Labs leaderboard Official board, date not stated |
| 6 | Muse Spark | Meta | 52.4 | Default | Scale Labs leaderboard Official board, date not stated |
| 7 | gpt-5 | OpenAI | 51.3 | Default | Scale Labs leaderboard Official board, date not stated |
| 8 | gpt-5-pro | OpenAI | 51.1 | Default | Scale Labs leaderboard Official board, date not stated |
| 9 | Fable 5.1 | Anthropic | 50.8 | Default | Scale Labs leaderboard Official board, date not stated |
| 10 | gpt-5.6-sol (max) | OpenAI | 50.5 | Default | Scale Labs leaderboard Official board, date not stated |
Papers and links
Last checked 1 Nov 2026. Something out of date or wrong? Send a correction.