OfficeQA
We show this table for reference; we do not rank on it.
Grounded numerical reasoning over a corpus of historical U.S. Treasury Bulletin documents.
Benchmark score on OfficeQA — November 13, 2026
We compile the OfficeQA rows from provider self-reports. Claude Fable 5.2 leads the table at 83.4%, followed by Claude Fable 5.1 (80.2%) and Claude Opus 5.5 (78.9%). We do not use these results to rank models overall.
Claude Fable 5.2
Anthropic
Reported by AnthropicClaude Fable 5.1
Anthropic
Reported by AnthropicClaude Opus 5.5
Anthropic
Reported by Anthropic5 modelsMultimodal & GroundedCurrentDisplay onlyUpdated November 13, 2026
Benchmark score table (5 models)
All 5 results on this page were reported by the labs that built the models.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Claude Fable 5.2AnthropicReported by Anthropic | 83.4% | Reported by Anthropic | Not reported | Closed |
| 2 | Claude Fable 5.1AnthropicReported by Anthropic | 80.2% | Reported by Anthropic | Not reported | Closed |
| 3 | Claude Opus 5.5AnthropicReported by Anthropic | 78.9% | Reported by Anthropic | Not reported | Closed |
| 4 | Claude Opus 5AnthropicReported by Anthropic | 78.1% | Reported by Anthropic | Not reported | Closed |
| 5 | Claude Sonnet 5.5AnthropicReported by Anthropic | 76.9% | Reported by Anthropic | Not reported | Closed |
- Lab-reported
- Benchmark-run
- Independent
Among the reported OfficeQA rows, Claude Fable 5.2 is first at 83.4%. The third row is 4.5 points behind. The broader top-10 range is 6.5 points, so many of the published results sit in a relatively narrow band.
5 models have been evaluated on OfficeQA. The benchmark falls in the Multimodal & Grounded category. OfficeQA is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About OfficeQA
Year
2026
Tasks
Historical Treasury Bulletin questions
Format
Agentic grounded QA accuracy
Difficulty
Professional document reasoning
Section 8.13.1 reports the full-set result using extracted documents and code execution. Opus 5 ran through the public Messages API with production safeguards.
Freshness and provenance
Version
OfficeQA 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 OfficeQA measure?
Grounded numerical reasoning over a corpus of historical U.S. Treasury Bulletin documents.
Which model scores highest on OfficeQA?
Claude Fable 5.2 by Anthropic currently leads with a score of 83.4% on OfficeQA.
How many models are evaluated on OfficeQA?
5 AI models have published results on OfficeQA in the BenchLM catalog.
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.