# Question Answering on HotPotQA (Success Rate and Best Delta)

Source: [Question Answering on HotPotQA (Success Rate and Best Delta)](<https://www.sota2.com/research/sota/question-answering-on-hotpotqa-success-rate-and-best-delta>)

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

Data updated: 2026-03-13T06:58:37.268Z

Evaluation of model performance on the HotpotQA dataset for Question Answering, measured using Success Rate (%) and Best Delta (pp).

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 | Question Answering |
| Dataset (including any stated split) | HotPotQA |
| Metric | Success Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 33 of 33 |
| 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

- [Success Rate](<https://www.sota2.com/research/sota/question-answering-on-hotpotqa-success-rate-and-best-delta.md?metric=Success%20Rate>): Higher is better; unit: PERCENT.

- [Best Delta (pp)](<https://www.sota2.com/research/sota/question-answering-on-hotpotqa-success-rate-and-best-delta.md?metric=Best%20Delta%20(pp)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Success Rate | Result date | Report paper |
| --- | --- | --- | --- |
| OPT | 82.7 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| LLaMA | 81 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Structured Agent Distillation | 75.2 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Structured Agent Distillation | 73.9 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Token-level baseline | 71.5 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| GLM-4.7 | 70.3 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| GLM-4.7 | 70.3 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| Token-level baseline | 69.7 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| SeqKD | 69.2 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| GLM-4.7 | 68.7 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| KD | 67.5 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| GLM-4.7 | 67.3 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| Structured Agent Distillation | 67 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| GLM-4.7 | 66.7 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| SeqKD | 66 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| KD | 64.3 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| MiniMax-M2.1 | 64 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| Token-level baseline | 62.9 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Structured Agent Distillation | 58.5 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| SeqKD | 58 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Qwen-3-8B | 57 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| KD | 56.2 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Qwen-3-8B | 54.7 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| Token-level baseline | 54.1 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Qwen-3-8B | 53.3 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| Qwen-3-8B | 51.3 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| SeqKD | 50.3 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| Qwen-3-8B | 49.7 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| KD | 49 | Not recorded | 68fa7bc61e209131cea8a5d4 |
| MiniMax-M2.1 | 38.5 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| MiniMax-M2.1 | 38 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| MiniMax-M2.1 | 36.7 | Not recorded | 6982d2a6a8b01e76cc8e9500 |
| MiniMax-M2.1 | 34.7 | Not recorded | 6982d2a6a8b01e76cc8e9500 |

### OPT

Success Rate: 82.7. Result ID: 69b3b59c47f34ce9249d2aab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Role":"Teacher","Size":"13B"}. This record does not assert independent reproduction.

### LLaMA

Success Rate: 81. Result ID: 69b3b59c47f34ce9249d2ab8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Role":"Teacher","Size":"13B"}. This record does not assert independent reproduction.

### Structured Agent Distillation

Success Rate: 75.2. Result ID: 69b3b59c47f34ce9249d2abc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"LLaMA-13B","Student Size":"7B"}. This record does not assert independent reproduction.

### Structured Agent Distillation

Success Rate: 73.9. Result ID: 69b3b59c47f34ce9249d2ab7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"6.7B"}. This record does not assert independent reproduction.

### Token-level baseline

Success Rate: 71.5. Result ID: 69b3b59c47f34ce9249d2abb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"LLaMA-13B","Student Size":"7B"}. This record does not assert independent reproduction.

### GLM-4.7

Success Rate: 70.3. Result ID: 699650455698665ec4950d91.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Baseline","Calibration":"N/A"}. This record does not assert independent reproduction.

### GLM-4.7

Success Rate: 70.3. Result ID: 699650455698665ec4950d92.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Rollback","Calibration":"Uncalibrated"}. This record does not assert independent reproduction.

### Token-level baseline

Success Rate: 69.7. Result ID: 69b3b59c47f34ce9249d2ab6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"6.7B"}. This record does not assert independent reproduction.

### SeqKD

Success Rate: 69.2. Result ID: 69b3b59c47f34ce9249d2aba.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62bb1e209131ce9faed1.md>)

[Implementation](<https://github.com/harvardnlp/seq2seq-attn>)

Recorded evaluation configuration: {"Teacher":"LLaMA-13B","Student Size":"7B"}. This record does not assert independent reproduction.

### GLM-4.7

Success Rate: 68.7. Result ID: 699650455698665ec4950d93.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Rollback","Calibration":"Calibrated"}. This record does not assert independent reproduction.

### KD

Success Rate: 67.5. Result ID: 69b3b59c47f34ce9249d2ab9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67221e209131cea14884.md>)

[Implementation](<https://github.com/huggingface/transformers>)

Recorded evaluation configuration: {"Teacher":"LLaMA-13B","Student Size":"7B"}. This record does not assert independent reproduction.

### GLM-4.7

Success Rate: 67.3. Result ID: 699650455698665ec4950d94.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Append","Calibration":"Uncalibrated"}. This record does not assert independent reproduction.

### Structured Agent Distillation

Success Rate: 67. Result ID: 69b3b59c47f34ce9249d2ab3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"2.7B"}. This record does not assert independent reproduction.

### GLM-4.7

Success Rate: 66.7. Result ID: 699650455698665ec4950d95.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Append","Calibration":"Calibrated"}. This record does not assert independent reproduction.

### SeqKD

Success Rate: 66. Result ID: 69b3b59c47f34ce9249d2ab5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62bb1e209131ce9faed1.md>)

[Implementation](<https://github.com/harvardnlp/seq2seq-attn>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"6.7B"}. This record does not assert independent reproduction.

### KD

Success Rate: 64.3. Result ID: 69b3b59c47f34ce9249d2ab4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67221e209131cea14884.md>)

[Implementation](<https://github.com/huggingface/transformers>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"6.7B"}. This record does not assert independent reproduction.

### MiniMax-M2.1

Success Rate: 64. Result ID: 699650455698665ec4950d96.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Baseline","Calibration":"N/A"}. This record does not assert independent reproduction.

### Token-level baseline

Success Rate: 62.9. Result ID: 69b3b59c47f34ce9249d2ab2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"2.7B"}. This record does not assert independent reproduction.

### Structured Agent Distillation

Success Rate: 58.5. Result ID: 69b3b59c47f34ce9249d2aaf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"1.3B"}. This record does not assert independent reproduction.

### SeqKD

Success Rate: 58. Result ID: 69b3b59c47f34ce9249d2ab1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62bb1e209131ce9faed1.md>)

[Implementation](<https://github.com/harvardnlp/seq2seq-attn>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"2.7B"}. This record does not assert independent reproduction.

### Qwen-3-8B

Success Rate: 57. Result ID: 699650455698665ec4950d8c.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

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

Recorded evaluation configuration: {"Intervention Mechanism":"Baseline","Calibration":"N/A"}. This record does not assert independent reproduction.

### KD

Success Rate: 56.2. Result ID: 69b3b59c47f34ce9249d2ab0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67221e209131cea14884.md>)

[Implementation](<https://github.com/huggingface/transformers>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"2.7B"}. This record does not assert independent reproduction.

### Qwen-3-8B

Success Rate: 54.7. Result ID: 699650455698665ec4950d8e.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

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

Recorded evaluation configuration: {"Intervention Mechanism":"Rollback","Calibration":"Calibrated"}. This record does not assert independent reproduction.

### Token-level baseline

Success Rate: 54.1. Result ID: 69b3b59c47f34ce9249d2aae.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"1.3B"}. This record does not assert independent reproduction.

### Qwen-3-8B

Success Rate: 53.3. Result ID: 699650455698665ec4950d8d.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

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

Recorded evaluation configuration: {"Intervention Mechanism":"Rollback","Calibration":"Uncalibrated"}. This record does not assert independent reproduction.

### Qwen-3-8B

Success Rate: 51.3. Result ID: 699650455698665ec4950d8f.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

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

Recorded evaluation configuration: {"Intervention Mechanism":"Append","Calibration":"Uncalibrated"}. This record does not assert independent reproduction.

### SeqKD

Success Rate: 50.3. Result ID: 69b3b59c47f34ce9249d2aad.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62bb1e209131ce9faed1.md>)

[Implementation](<https://github.com/harvardnlp/seq2seq-attn>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"1.3B"}. This record does not assert independent reproduction.

### Qwen-3-8B

Success Rate: 49.7. Result ID: 699650455698665ec4950d90.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

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

Recorded evaluation configuration: {"Intervention Mechanism":"Append","Calibration":"Calibrated"}. This record does not assert independent reproduction.

### KD

Success Rate: 49. Result ID: 69b3b59c47f34ce9249d2aac.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67221e209131cea14884.md>)

[Implementation](<https://github.com/huggingface/transformers>)

Recorded evaluation configuration: {"Teacher":"OPT-13B","Student Size":"1.3B"}. This record does not assert independent reproduction.

### MiniMax-M2.1

Success Rate: 38.5. Result ID: 699650455698665ec4950d98.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Rollback","Calibration":"Calibrated"}. This record does not assert independent reproduction.

### MiniMax-M2.1

Success Rate: 38. Result ID: 699650455698665ec4950d97.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Rollback","Calibration":"Uncalibrated"}. This record does not assert independent reproduction.

### MiniMax-M2.1

Success Rate: 36.7. Result ID: 699650455698665ec4950d9a.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Append","Calibration":"Calibrated"}. This record does not assert independent reproduction.

### MiniMax-M2.1

Success Rate: 34.7. Result ID: 699650455698665ec4950d99.

[Reporting paper](<https://www.sota2.com/research/paper/6982d2a6a8b01e76cc8e9500.md>)

Recorded evaluation configuration: {"Intervention Mechanism":"Append","Calibration":"Uncalibrated"}. This record does not assert independent reproduction.

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