OfficeQA Pro
A benchmark for grounded reasoning over office-style documents, spreadsheets, charts, and business artifacts.
Top models on OfficeQA Pro — November 13, 2026
As of November 13, 2026, Claude Fable 5.2 leads the OfficeQA Pro leaderboard with 72.9% , followed by Claude Fable 5.1 (69%) and Claude Opus 5.5 (67.7%).
Claude Fable 5.2
Anthropic
Reported by AnthropicClaude Fable 5.1
Anthropic
Reported by AnthropicClaude Opus 5.5
Anthropic
Reported by Anthropic16 modelsMultimodal & Grounded20% of Multimodal reference weightCurrentUpdated November 13, 2026
Leaderboard (16 models)
None of the 16 results on this page has a documented run by the benchmark or an independent evaluator: 15 reported by the labs that built the models and 1 reported in another publisher’s document.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Claude Fable 5.2AnthropicReported by Anthropic | 72.9% | Reported by Anthropic | Not reported | Closed |
| 2 | Claude Fable 5.1AnthropicReported by Anthropic | 69% | Reported by Anthropic | Not reported | Closed |
| 3 | Claude Opus 5.5AnthropicReported by Anthropic | 67.7% | Reported by Anthropic | Not reported | Closed |
| 4 | Kimi K3.1 ProMoonshot AIReported by Moonshot AI | 67.4% | Reported by Moonshot AI | Not reported | Open |
| 5 | Claude Opus 5AnthropicReported by Anthropic | 66.9% | Reported by Anthropic | Not reported | Closed |
| 6 | Claude Opus 4.8AnthropicReported by Anthropic | 66.2% | Reported by Anthropic | Not reported | Closed |
| 7 | Hy4 previewTencentReported by Tencent | 66.2% | Reported by Tencent | Not reported | Open |
| 8 | Claude Sonnet 5.5AnthropicReported by Anthropic | 65.6% | Reported by Anthropic | Not reported | Closed |
| 9 | Kimi K3Moonshot AIReported by Moonshot AI | 63.3% | Reported by Moonshot AI | Not reported | Open |
| 10 | GLM-5.3-FlashZ.AIReported by Z.AI | 62.4% | Reported by Z.AI | Not reported | Open |
| 11 | Step 5 PreviewStepFunReported by StepFun | 60.3% | Reported by StepFun | Not reported | Open |
| 12 | Claude Fable 5AnthropicReported by Anthropic | 57.9% | Reported by Anthropic | Not reported | Closed |
| 13 | GPT-5.5OpenAIReported by OpenAI | 54.1% | Reported by OpenAI | Not reported | Closed |
| 14 | GPT-5.4OpenAIReported by OpenAI | 53.2% | Reported by OpenAI | Not reported | Closed |
| 15 | MiniMax M3MiniMaxReported by MiniMax | 45.1% | Reported by MiniMax | Not reported | Open |
| 16 | Claude Opus 4.7 (Adaptive)AnthropicReported by OpenAI (comparison table) | 43.6% | Reported by OpenAI (comparison table) | Not reported | Closed |
- Lab-reported
- Benchmark-run
- Independent
OfficeQA Pro v2
Moonshot evaluates with Claude Code (Codex for GPT models), renders every PDF in the corpus as images with no machine-readable text, and scores zero for tasks where a model retrieved reference answers online. BenchLM keeps v2 separate from the original OfficeQA Pro key.
| Source rank | Source model or configuration | Score | Who ran it | Source setup | Parameters (B) | Open / closed |
|---|---|---|---|---|---|---|
| 1 | ERNIE 5.5 Flash PreviewBaidu | 67.9% | Reported by Baidu | Moonshot AIOriginal result | Not reported | Pending |
| 2 | Kimi K3.1 ProMoonshot AI | 64.4% | Reported by Moonshot AI | Moonshot AIOriginal result | Not reported | Open |
According to BenchLM.ai, Claude Fable 5.2 leads the OfficeQA Pro benchmark with a score of 72.9%, followed by Claude Fable 5.1 (69%) and Claude Opus 5.5 (67.7%). The scores show moderate spread, with meaningful differences between the top tier and mid-tier models.
16 models have been evaluated on OfficeQA Pro. The benchmark falls in the Multimodal & Grounded category. BenchAlign v5.8 gives OfficeQA Pro 20% of the Multimodal reference weight, so it moves the Multimodal leaderboard. Reference weights are relative weights in the calibrated model, not fixed shares of a score.
About OfficeQA Pro
Year
2026
Tasks
Document and spreadsheet tasks
Format
Grounded QA over office artifacts
Difficulty
Enterprise grounded reasoning
OfficeQA Pro is useful when choosing models for enterprise copilots because it measures whether they can reason correctly over real office content rather than generic chat prompts.
Freshness and provenance
Version
OfficeQA Pro 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 Pro measure?
A benchmark for grounded reasoning over office-style documents, spreadsheets, charts, and business artifacts.
Which model scores highest on OfficeQA Pro?
Claude Fable 5.2 by Anthropic currently leads with a score of 72.9% on OfficeQA Pro.
How many models are evaluated on OfficeQA Pro?
16 AI models have published results on OfficeQA Pro in the BenchLM catalog.
Compare top models on OfficeQA Pro
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.