# Long-term Memory Question Answering on ATM-Bench 1.0 (test)

Source: [Long-term Memory Question Answering on ATM-Bench 1.0 (test)](<https://www.sota2.com/research/sota/long-term-memory-question-answering-on-atm-bench-1-0-test>)

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

Data updated: 2026-05-06T08:46:59.347Z

Evaluation of model performance on the test split of ATM-Bench version 1.0, focusing on Long-term Memory Question Answering capabilities, measured by metrics including QS, R@10, and Joint@10.

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 | Long-term Memory Question Answering |
| Dataset (including any stated split) | ATM-Bench 1.0 (test) |
| Metric | QS |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [QS](<https://www.sota2.com/research/sota/long-term-memory-question-answering-on-atm-bench-1-0-test.md?metric=QS>): Higher is better; unit: PERCENT.

- [R@10](<https://www.sota2.com/research/sota/long-term-memory-question-answering-on-atm-bench-1-0-test.md?metric=R%4010>): Higher is better; unit: PERCENT.

- [Joint@10](<https://www.sota2.com/research/sota/long-term-memory-question-answering-on-atm-bench-1-0-test.md?metric=Joint%4010>): Higher is better; unit: PERCENT.

- [N](<https://www.sota2.com/research/sota/long-term-memory-question-answering-on-atm-bench-1-0-test.md?metric=N>): Higher is better; unit: PERCENT.

- [R](<https://www.sota2.com/research/sota/long-term-memory-question-answering-on-atm-bench-1-0-test.md?metric=R>): Higher is better; unit: PERCENT.

- [O](<https://www.sota2.com/research/sota/long-term-memory-question-answering-on-atm-bench-1-0-test.md?metric=O>): Higher is better; unit: PERCENT.

## Recorded results

| Method | QS | Result date | Report paper |
| --- | --- | --- | --- |
| Oracle | 77.8 | Not recorded | 69facb18c0f56ed84308a404 |
| Oracle | 70 | Not recorded | 69facb18c0f56ed84308a404 |
| ScrapMem (No-Forget) | 52.5 | Not recorded | 69facb18c0f56ed84308a404 |
| ATM-RAG | 51 | Not recorded | 69facb18c0f56ed84308a404 |
| Self-RAG | 50.3 | Not recorded | 69facb18c0f56ed84308a404 |
| ScrapMem | 48.4 | Not recorded | 69facb18c0f56ed84308a404 |
| HippoRAG2 | 47.7 | Not recorded | 69facb18c0f56ed84308a404 |
| A-Mem (Piled) | 46.1 | Not recorded | 69facb18c0f56ed84308a404 |
| A-Mem (Linked) | 44.8 | Not recorded | 69facb18c0f56ed84308a404 |
| Mem0Agentic | 43.5 | Not recorded | 69facb18c0f56ed84308a404 |
| HippoRAG2 | 42.9 | Not recorded | 69facb18c0f56ed84308a404 |
| Self-RAG | 42.1 | Not recorded | 69facb18c0f56ed84308a404 |
| ATM-RAG | 42 | Not recorded | 69facb18c0f56ed84308a404 |
| Mem0Plain | 41.5 | Not recorded | 69facb18c0f56ed84308a404 |
| No-Evidence | 0.2 | Not recorded | 69facb18c0f56ed84308a404 |

### Oracle

QS: 77.8. Result ID: 69fb000047f34ce924de010f.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

Recorded evaluation configuration: {"Memory Representation":"SGM"}. This record does not assert independent reproduction.

### Oracle

QS: 70. Result ID: 69fb000047f34ce924de010e.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### ScrapMem (No-Forget)

QS: 52.5. Result ID: 69fb000047f34ce924de011a.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

Recorded evaluation configuration: {"Memory Representation":"Graph"}. This record does not assert independent reproduction.

### ATM-RAG

QS: 51. Result ID: 69fb000047f34ce924de0119.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0da.md>)

[Implementation](<https://github.com/JingbiaoMei/ATM-Bench>)

Recorded evaluation configuration: {"Memory Representation":"SGM"}. This record does not assert independent reproduction.

### Self-RAG

QS: 50.3. Result ID: 69fb000047f34ce924de0117.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

Recorded evaluation configuration: {"Memory Representation":"SGM"}. This record does not assert independent reproduction.

### ScrapMem

QS: 48.4. Result ID: 69fb000047f34ce924de011b.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

Recorded evaluation configuration: {"Memory Representation":"Graph"}. This record does not assert independent reproduction.

### HippoRAG2

QS: 47.7. Result ID: 69fb000047f34ce924de0115.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79821e209131cea7de8a.md>)

Recorded evaluation configuration: {"Memory Representation":"SGM"}. This record does not assert independent reproduction.

### A-Mem (Piled)

QS: 46.1. Result ID: 69fb000047f34ce924de0110.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bf31e209131cea8b869.md>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### A-Mem (Linked)

QS: 44.8. Result ID: 69fb000047f34ce924de0111.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bf31e209131cea8b869.md>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### Mem0Agentic

QS: 43.5. Result ID: 69fb000047f34ce924de0112.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa784b1e209131cea77ad6.md>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### HippoRAG2

QS: 42.9. Result ID: 69fb000047f34ce924de0114.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79821e209131cea7de8a.md>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### Self-RAG

QS: 42.1. Result ID: 69fb000047f34ce924de0116.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### ATM-RAG

QS: 42. Result ID: 69fb000047f34ce924de0118.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0da.md>)

[Implementation](<https://github.com/JingbiaoMei/ATM-Bench>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### Mem0Plain

QS: 41.5. Result ID: 69fb000047f34ce924de0113.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa784b1e209131cea77ad6.md>)

Recorded evaluation configuration: {"Memory Representation":"DM"}. This record does not assert independent reproduction.

### No-Evidence

QS: 0.2. Result ID: 69fb000047f34ce924de010d.

[Reporting paper](<https://www.sota2.com/research/paper/69facb18c0f56ed84308a404.md>)

Recorded evaluation configuration: {"Memory Representation":"–"}. This record does not assert independent reproduction.

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