SpreadsheetBench 2
SpreadsheetBench 2 - agent evaluation on end-to-end spreadsheet tasks, NeurIPS 2026 successor of SpreadsheetBench. Scores are pass rates over the SpreadsheetBench-v2 dataset (https://huggingface.co/datasets/KAKA22/SpreadsheetBench-v2); code: https://github.com/RUCKBReasoning/SpreadsheetBench-2.
Category
General Agents
Score type
percent
Scale
100%
higher is better
Models tested
3
What is SpreadsheetBench 2?
SpreadsheetBench 2 is a benchmark for how well a model acts as a general-purpose agent. The official description says: "SpreadsheetBench 2 - agent evaluation on end-to-end spreadsheet tasks, NeurIPS 2026 successor of SpreadsheetBench". Below we describe in simple words what benchmarks of this type generally examine.
What does it test?
SpreadsheetBench 2 measures how well a model acts as a general-purpose agent. In detail, it checks:
- solving long, multi-step tasks with many intermediate decisions
- using tools correctly (browsers, terminals, APIs, calendars, other apps)
- adapting when something unexpected happens mid-task
- staying on goal over long task horizons instead of wandering off
How to read the score
The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.
Official methodology note: SpreadsheetBench 2 - agent evaluation on end-to-end spreadsheet tasks, NeurIPS 2026 successor of SpreadsheetBench. Scores are pass rates over the SpreadsheetBench-v2 dataset (https://huggingface.co/datasets/KAKA22/SpreadsheetBench-v2); code: https://github.com/RUCKBReasoning/SpreadsheetBench-2.
Frequently asked questions
What is the SpreadsheetBench 2 benchmark?
SpreadsheetBench 2 is a benchmark for how well a model acts as a general-purpose agent. The official description says: 'SpreadsheetBench 2 - agent evaluation on end-to-end spreadsheet tasks, NeurIPS 2026 successor of SpreadsheetBench'. Below we describe in simple words what benchmarks of this type generally examine.
What does SpreadsheetBench 2 test?
SpreadsheetBench 2 mainly measures how well a model acts as a general-purpose agent. Concretely, it checks: solving long, multi-step tasks with many intermediate decisions; using tools correctly (browsers, terminals, APIs, calendars, other apps); adapting when something unexpected happens mid-task.
How is the SpreadsheetBench 2 score calculated?
The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better. The declared maximum score of this benchmark is 100.0.
Where do I get SpreadsheetBench 2 data and scores?
The official source is https://huggingface.co/datasets/KAKA22/SpreadsheetBench-v2. The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/spreadsheetbench-2/scores) or pull it directly from the Open LLM Radar API with any HTTP client.
Benchmark data & downloads
Where to get, inspect and reuse the data behind this benchmark.
Official benchmark page / paper
The benchmark maintainer's website, paper or leaderboard
Benchmark dataset
The question/task data itself — download and evaluate locally
Benchmark metadata (JSON)
This page's benchmark record from the Open LLM Radar API
All model scores (JSON download)
Full score table as JSON; paginate with limit/offset params
Open LLM Radar API docs
Swagger UI to explore and reuse all radar API endpoints
Model scores 3 models tested
| # | Model | Score | Date | Source | Verified |
|---|---|---|---|---|---|
| 1 |
Sage-S2.1-Flash
Shanghai AI Laboratory
|
37.1%
|
— | Report | Verified |
| 2 |
DeepSeek-V4.1-Flash-1017
DeepSeek
|
35.8%
|
— | Report | Unverified |
| 3 |
GLM-5.3-Flash
Zhipu AI
|
32.4%
|
— | Report | Unverified |