Spreadsheet Understanding on FRTR-Bench
99AccuracyBRTR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| BRTRModel=GPT-5, Embedding Model=NVIDIA NeMo Retriever2026.03 | 99 | 4.01 | 68,061 | 166.34 | |
| BRTRModel=Gemini 3 Pro, Embedding Model=NVIDIA NeMo Retriever2026.03 | 99 | 5.4 | 48,462 | 217.77 | |
| BRTRModel=Claude Opus 4.6, Embedding Model=NVIDIA NeMo Retriever2026.03 | 98 | 5.2 | 462,614 | 143.23 | |
| BRTRModel=GPT-5.2, Embedding Model=NVIDIA NeMo Retriever2026.03 | 98 | 3.05 | 20,127 | 90.81 | |
| BRTRModel=Claude Sonnet 4.5, Embedding Model=NVIDIA NeMo Retriever2026.03 | 96 | 6.2 | 92,202 | 163.83 | |
| BRTRModel=Gemini 2.5 Pro, Embedding Model=NVIDIA NeMo Retriever2026.03 | 85 | 2.23 | 15,729 | 111.04 | |
| Claude in ExcelModel=Claude Opus 4.62026.03 | 84 | — | — | — | |
| FRTRModel=Claude Opus 4.6, Mean Context Size=7,691 tokens2026.03 | 74 | — | — | — | |
| FRTRModel=Claude Sonnet 4.5, Mean Context Size=7,691 tokens2026.03 | 74 | — | — | — | |
| FRTRModel=GPT-5, Mean Context Size=7,691 tokens2026.03 | 73 | — | — | — | |
| FRTRModel=GPT-5.2, Mean Context Size=7,691 tokens2026.03 | 72 | — | — | — | |
| Claude in ExcelModel=Claude Sonnet 4.52026.03 | 69 | — | — | — | |
| FRTRModel=Gemini 3 Pro, Mean Context Size=7,691 tokens2026.03 | 68 | — | — | — | |
| BRTRModel=GPT-4o, Embedding Model=NVIDIA NeMo Retriever2026.03 | 67 | 3.25 | 50,745 | 140.2 | |
| BRTRModel=LLaMA 4 Maverick 17B, Embedding Model=NVIDIA NeMo Retriever2026.03 | 67 | 1.99 | 38,483 | 92.32 | |
| FRTRModel=Gemini 2.5 Pro, Mean Context Size=7,691 tokens2026.03 | 67 | — | — | — | |
| BRTRModel=Pixtral Large 2502, Embedding Model=NVIDIA NeMo Retriever2026.03 | 63 | 3.12 | 70,618 | 51.76 | |
| FRTRModel=LLaMA 4 Maverick 17B, Mean Context Size=7,691 tokens2026.03 | 56 | — | — | — | |
| FRTRModel=GPT-4o, Mean Context Size=7,691 tokens2026.03 | 49 | — | — | — | |
| FRTRModel=Pixtral Large 2502, Mean Context Size=7,691 tokens2026.03 | 47 | — | — | — | |
| SpreadsheetLLMModel=Claude Opus 4.6, Mean Context Size=12,745 tokens2026.03 | 34 | — | — | — | |
| SpreadsheetLLMModel=GPT-5.2, Mean Context Size=12,745 tokens2026.03 | 32 | — | — | — | |
| SpreadsheetLLMModel=Gemini 3 Pro, Mean Context Size=12,745 tokens2026.03 | 29 | — | — | — | |
| SpreadsheetLLMModel=Gemini 2.5 Pro, Mean Context Size=12,745 tokens2026.03 | 24 | — | — | — | |
| SpreadsheetLLMModel=LLaMA 4 Maverick 17B, Mean Context Size=12,745 tokens2026.03 | 23 | — | — | — | |
| SpreadsheetLLMModel=Pixtral Large 2502, Mean Context Size=12,745 tokens2026.03 | 21 | — | — | — | |
| SpreadsheetLLMModel=GPT-5, Mean Context Size=12,745 tokens2026.03 | 18 | — | — | — | |
| SpreadsheetLLMModel=Claude Sonnet 4.5, Mean Context Size=12,745 tokens2026.03 | 13 | — | — | — | |
| SpreadsheetLLMModel=GPT-4o, Mean Context Size=12,745 tokens2026.03 | 6 | — | — | — |