# General AI Assistant on GAIA text

Source: [General AI Assistant on GAIA text](<https://www.sota2.com/research/sota/general-ai-assistant-on-gaia-text>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:45:39.662Z

Evaluation of general AI assistant capabilities on the text-only version of the GAIA dataset, measured by Level 1, Level 2, Level 3, and Average scores.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | General AI Assistant |
| Dataset (including any stated split) | GAIA text |
| Metric | GAIA Average Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 19 of 19 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [GAIA Average Score](<https://www.sota2.com/research/sota/general-ai-assistant-on-gaia-text.md?metric=GAIA%20Average%20Score>): Higher is better; unit: PERCENT.

- [GAIA Level 1 Score](<https://www.sota2.com/research/sota/general-ai-assistant-on-gaia-text.md?metric=GAIA%20Level%201%20Score>): Higher is better; unit: PERCENT.

- [GAIA Level 2 Score](<https://www.sota2.com/research/sota/general-ai-assistant-on-gaia-text.md?metric=GAIA%20Level%202%20Score>): Higher is better; unit: PERCENT.

- [GAIA Level 3 Score](<https://www.sota2.com/research/sota/general-ai-assistant-on-gaia-text.md?metric=GAIA%20Level%203%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | GAIA Average Score | Result date | Report paper |
| --- | --- | --- | --- |
| OpenAI-o3 | 70.5 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Claude-4-Sonnet | 68.3 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| OpenAI DeepResearch | 67.4 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| ARPO | 43.7 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Reagent-U | 43.7 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| WebDancer | 40.7 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Search-o1 | 39.8 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| ARPO | 38.8 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Reagent-R | 36.9 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| VerlTool | 34 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Reagent w/o Agent-RRM | 34 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| WebDancer | 31 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| DeepSeek-R1-671B | 25.2 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Reagent-C | 25.2 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| WebThinker | 22.3 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Qwen3-8B | 21.4 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Tree-GRPO | 21 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| QwQ-32B | 18.9 | Not recorded | 697c3b40a8b01e76cc8da5d5 |
| Atom-Searcher | 17.5 | Not recorded | 697c3b40a8b01e76cc8da5d5 |

### OpenAI-o3

GAIA Average Score: 70.5. Result ID: 699616c15698665ec45668d7.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

Recorded evaluation configuration: {"Backbone":null}. This record does not assert independent reproduction.

### Claude-4-Sonnet

GAIA Average Score: 68.3. Result ID: 699616c15698665ec45668d8.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

Recorded evaluation configuration: {"Backbone":null}. This record does not assert independent reproduction.

### OpenAI DeepResearch

GAIA Average Score: 67.4. Result ID: 699616c15698665ec45668d9.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

Recorded evaluation configuration: {"Backbone":null}. This record does not assert independent reproduction.

### ARPO

GAIA Average Score: 43.7. Result ID: 699616c15698665ec45668e1.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a331e209131cea81b93.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-14B"}. This record does not assert independent reproduction.

### Reagent-U

GAIA Average Score: 43.7. Result ID: 699616c15698665ec45668e9.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Implementation](<https://github.com/kxfan2002/Reagent>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### WebDancer

GAIA Average Score: 40.7. Result ID: 699616c15698665ec45668e3.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a8d1e209131cea83dce.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-32B"}. This record does not assert independent reproduction.

### Search-o1

GAIA Average Score: 39.8. Result ID: 699616c15698665ec45668e2.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76911e209131cea6dca9.md>)

[Implementation](<https://github.com/sunnynexus/Search-o1>)

Recorded evaluation configuration: {"Backbone":"QwQ-32B-Preview"}. This record does not assert independent reproduction.

### ARPO

GAIA Average Score: 38.8. Result ID: 699616c15698665ec45668dd.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a331e209131cea81b93.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Reagent-R

GAIA Average Score: 36.9. Result ID: 699616c15698665ec45668e8.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Implementation](<https://github.com/kxfan2002/Reagent>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### VerlTool

GAIA Average Score: 34. Result ID: 699616c15698665ec45668dc.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c441e209131cea8d468.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Reagent w/o Agent-RRM

GAIA Average Score: 34. Result ID: 699616c15698665ec45668e6.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Implementation](<https://github.com/kxfan2002/Reagent>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### WebDancer

GAIA Average Score: 31. Result ID: 699616c15698665ec45668db.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a8d1e209131cea83dce.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B"}. This record does not assert independent reproduction.

### DeepSeek-R1-671B

GAIA Average Score: 25.2. Result ID: 699616c15698665ec45668df.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76a91e209131cea6e3eb.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-671B"}. This record does not assert independent reproduction.

### Reagent-C

GAIA Average Score: 25.2. Result ID: 699616c15698665ec45668e7.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Implementation](<https://github.com/kxfan2002/Reagent>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Inference Type":"Direct Inference"}. This record does not assert independent reproduction.

### WebThinker

GAIA Average Score: 22.3. Result ID: 699616c15698665ec45668da.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c191e209131cea8c4d7.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Qwen3-8B

GAIA Average Score: 21.4. Result ID: 699616c15698665ec45668e5.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Tree-GRPO

GAIA Average Score: 21. Result ID: 699616c15698665ec45668e0.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c349.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-14B"}. This record does not assert independent reproduction.

### QwQ-32B

GAIA Average Score: 18.9. Result ID: 699616c15698665ec45668de.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

Recorded evaluation configuration: {"Backbone":"QwQ-32B"}. This record does not assert independent reproduction.

### Atom-Searcher

GAIA Average Score: 17.5. Result ID: 699616c15698665ec45668e4.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b40a8b01e76cc8da5d5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7af01e209131cea85dba.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
