Anthropic Real-World Finance v2 (Real-World Finance v2)
We show this table for reference; we do not rank on it.
Anthropic's internal evaluation of end-to-end financial-analyst work: 123 tasks, each model's attempts compared pairwise by a grader and rated on an Elo scale.
Benchmark score on Real-World Finance v2 — November 11, 2026
We compile the Real-World Finance v2 rows from provider self-reports. Claude Fable 5.2 leads the table at 1637, followed by Claude Opus 5.5 (1571) and Claude Fable 5.1 (1493). We do not use these results to rank models overall.
Claude Fable 5.2
Anthropic
Claude Opus 5.5
Anthropic
Claude Fable 5.1
Anthropic
3 modelsAgenticCurrentDisplay onlyUpdated November 11, 2026
Benchmark score table (3 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Claude Fable 5.2Anthropic | 1637 | Not reported | Closed |
| 2 | Claude Opus 5.5Anthropic | 1571 | Not reported | Closed |
| 3 | Claude Fable 5.1Anthropic | 1493 | Not reported | Closed |
Among the reported Real-World Finance v2 rows, Claude Fable 5.2 is first at 1637. The third row is 144 score units behind. The broader top-10 range is 144 score units, so the table still separates the published systems.
3 models have been evaluated on Real-World Finance v2. The benchmark falls in the Agentic category. Real-World Finance v2 is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Real-World Finance v2
Year
2026
Tasks
123 finance tasks
Format
Elo (pairwise grader)
Difficulty
Professional finance work
Since the Fable 5.2 system card the grader is built on Claude Opus 5.5 and models get more time and context; Anthropic divides the new grader's Elo gaps by 2.61 to stay on the scale of earlier system cards, where Claude Sonnet 4.6 = 1000. Internal and Elo-scale, so display-only and outside the Data catalog.
Freshness and provenance
Version
Real-World Finance v2 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Real-World Finance v2 measure?
Anthropic's internal evaluation of end-to-end financial-analyst work: 123 tasks, each model's attempts compared pairwise by a grader and rated on an Elo scale.
Which model scores highest on Real-World Finance v2?
Claude Fable 5.2 by Anthropic currently leads with a score of 1637 on Real-World Finance v2.
How many models are evaluated on Real-World Finance v2?
3 AI models have published results on Real-World Finance v2 in the BenchLM catalog.
Compare top models on Real-World Finance v2
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.