# LLM Agent Reasoning on BrowseComp-Plus

Source: [LLM Agent Reasoning on BrowseComp-Plus](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus>)

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

Data updated: 2026-03-10T11:22:02.363Z

Evaluation of LLM Agent Reasoning capabilities on the BrowseComp-Plus dataset, focusing on accuracy and resource efficiency metrics such as total tokens, task tokens, and step counts.

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 | LLM Agent Reasoning |
| Dataset (including any stated split) | BrowseComp-Plus |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Change in Accuracy](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Change%20in%20Accuracy>): Higher is better; unit: SCALAR.

- [Score](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Score>): Higher is better; unit: SCALAR.

- [Total Tokens](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Total%20Tokens>): Lower is better; unit: SIZE.

- [Change in Total Tokens](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Change%20in%20Total%20Tokens>): Lower is better; unit: SIZE.

- [Task Tokens](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Task%20Tokens>): Lower is better; unit: SIZE.

- [Change in Task Tokens](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Change%20in%20Task%20Tokens>): Lower is better; unit: SIZE.

- [Total Steps](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Total%20Steps>): Lower is better; unit: SCALAR.

- [Change in Step Tokens](<https://www.sota2.com/research/sota/llm-agent-reasoning-on-browsecomp-plus.md?metric=Change%20in%20Step%20Tokens>): Lower is better; unit: SIZE.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Rule-Based (High) | 42.7 | Not recorded | 69afa5932d603e6de3e43ba8 |
| ARES | 41.3 | Not recorded | 69afa5932d603e6de3e43ba8 |
| Prompting-Based (GPT 5) | 38.7 | Not recorded | 69afa5932d603e6de3e43ba8 |
| Prompting-Based (Gemini 3 Pro) | 37.3 | Not recorded | 69afa5932d603e6de3e43ba8 |
| Rule-Based (Medium) | 34 | Not recorded | 69afa5932d603e6de3e43ba8 |
| Rule-Based (Random) | 30.7 | Not recorded | 69afa5932d603e6de3e43ba8 |
| Rule-Based (Low) | 8 | Not recorded | 69afa5932d603e6de3e43ba8 |

### Rule-Based (High)

Accuracy: 42.7. Result ID: 69affeca6dbb7075b95967e0.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

Recorded evaluation configuration: {"Strategy":"Rule-Based","Effort Level":"High","Backbone LLM":"gpt-oss-20b"}. This record does not assert independent reproduction.

### ARES

Accuracy: 41.3. Result ID: 69affeca6dbb7075b95967e4.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

Recorded evaluation configuration: {"Strategy":"Adaptive Selection","Backbone LLM":"gpt-oss-20b"}. This record does not assert independent reproduction.

### Prompting-Based (GPT 5)

Accuracy: 38.7. Result ID: 69affeca6dbb7075b95967e2.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/695f39f053fdd5c361743ecb.md>)

Recorded evaluation configuration: {"Strategy":"Prompting-Based","Model":"GPT 5","Backbone LLM":"gpt-oss-20b"}. This record does not assert independent reproduction.

### Prompting-Based (Gemini 3 Pro)

Accuracy: 37.3. Result ID: 69affeca6dbb7075b95967e3.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

Recorded evaluation configuration: {"Strategy":"Prompting-Based","Model":"Gemini 3 Pro","Backbone LLM":"gpt-oss-20b"}. This record does not assert independent reproduction.

### Rule-Based (Medium)

Accuracy: 34. Result ID: 69affeca6dbb7075b95967df.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

Recorded evaluation configuration: {"Strategy":"Rule-Based","Effort Level":"Medium","Backbone LLM":"gpt-oss-20b"}. This record does not assert independent reproduction.

### Rule-Based (Random)

Accuracy: 30.7. Result ID: 69affeca6dbb7075b95967e1.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

Recorded evaluation configuration: {"Strategy":"Rule-Based","Effort Level":"Random","Backbone LLM":"gpt-oss-20b"}. This record does not assert independent reproduction.

### Rule-Based (Low)

Accuracy: 8. Result ID: 69affeca6dbb7075b95967de.

[Reporting paper](<https://www.sota2.com/research/paper/69afa5932d603e6de3e43ba8.md>)

Recorded evaluation configuration: {"Strategy":"Rule-Based","Effort Level":"Low","Backbone LLM":"gpt-oss-20b"}. This record does not assert independent reproduction.

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