# Discourse-level Chinese-English translation on DiscoX Source: [Discourse-level Chinese-English translation on DiscoX]() Collection index: [llms.txt]() Data updated: 2026-04-06T08:44:15.155Z Evaluation of model performance on discourse-level Chinese-English translation using the DiscoX dataset, measured by Accuracy, Fluency, and Approximation. 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 | Discourse-level Chinese-English translation | | Dataset (including any stated split) | DiscoX | | Metric | Accuracy | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [Fluency](): Higher is better; unit: SCALAR. - [Approximation](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | GPT-5 | 42.62 | Not recorded | 69d33ded7d0172140ec735d6 | | Gemini2.5-Pro | 40.23 | Not recorded | 69d33ded7d0172140ec735d6 | | Claude-4-sonnet | 38.65 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + R2-Write-SFT + RLp | 23.5 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + R2-Write-SFT + RL | 23.16 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + R2-Write-SFT | 22.7 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + Distill | 21.7 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + Distill + RL | 20.42 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + R2-Write-Last | 20.14 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + RL | 19.3 | Not recorded | 69d33ded7d0172140ec735d6 | | LongWriter-Zero-32B | 19.2 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B | 18.8 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B + R2-Write-SFT + RLp | 18.64 | Not recorded | 69d33ded7d0172140ec735d6 | | Reverse-Engineering-8B | 18.62 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B + R2-Write-SFT | 18.22 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B + RL | 15.26 | Not recorded | 69d33ded7d0172140ec735d6 | | Writing-RL-7B | 15.2 | Not recorded | 69d33ded7d0172140ec735d6 | | Longwriter-9B | 14.12 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B | 12.82 | Not recorded | 69d33ded7d0172140ec735d6 | ### GPT-5 Accuracy: 42.62. Result ID: 69d3725e47f34ce924bb5c95. [Reporting paper]() Recorded evaluation configuration: {"Training":"-"}. This record does not assert independent reproduction. ### Gemini2.5-Pro Accuracy: 40.23. Result ID: 69d3725e47f34ce924bb5c96. [Reporting paper]() Recorded evaluation configuration: {"Training":"-"}. This record does not assert independent reproduction. ### Claude-4-sonnet Accuracy: 38.65. Result ID: 69d3725e47f34ce924bb5c97. [Reporting paper]() Recorded evaluation configuration: {"Training":"-"}. This record does not assert independent reproduction. ### Qwen3-8B + R2-Write-SFT + RLp Accuracy: 23.5. Result ID: 69d3725e47f34ce924bb5ca7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + PPO"}. This record does not assert independent reproduction. ### Qwen3-8B + R2-Write-SFT + RL Accuracy: 23.16. Result ID: 69d3725e47f34ce924bb5ca6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + PPO"}. This record does not assert independent reproduction. ### Qwen3-8B + R2-Write-SFT Accuracy: 22.7. Result ID: 69d3725e47f34ce924bb5c9f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-8B + Distill Accuracy: 21.7. Result ID: 69d3725e47f34ce924bb5c9c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-8B + Distill + RL Accuracy: 20.42. Result ID: 69d3725e47f34ce924bb5ca2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + PPO"}. This record does not assert independent reproduction. ### Qwen3-8B + R2-Write-Last Accuracy: 20.14. Result ID: 69d3725e47f34ce924bb5c9d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-8B + RL Accuracy: 19.3. Result ID: 69d3725e47f34ce924bb5ca1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"PPO"}. This record does not assert independent reproduction. ### LongWriter-Zero-32B Accuracy: 19.2. Result ID: 69d3725e47f34ce924bb5ca4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + GRPO"}. This record does not assert independent reproduction. ### Qwen3-8B Accuracy: 18.8. Result ID: 69d3725e47f34ce924bb5c99. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"-"}. This record does not assert independent reproduction. ### Qwen3-4B + R2-Write-SFT + RLp Accuracy: 18.64. Result ID: 69d3725e47f34ce924bb5ca5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + PPO"}. This record does not assert independent reproduction. ### Reverse-Engineering-8B Accuracy: 18.62. Result ID: 69d3725e47f34ce924bb5c9b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-4B + R2-Write-SFT Accuracy: 18.22. Result ID: 69d3725e47f34ce924bb5c9e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-4B + RL Accuracy: 15.26. Result ID: 69d3725e47f34ce924bb5ca0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"PPO"}. This record does not assert independent reproduction. ### Writing-RL-7B Accuracy: 15.2. Result ID: 69d3725e47f34ce924bb5ca3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + PPO"}. This record does not assert independent reproduction. ### Longwriter-9B Accuracy: 14.12. Result ID: 69d3725e47f34ce924bb5c9a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-4B Accuracy: 12.82. Result ID: 69d3725e47f34ce924bb5c98. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"-"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.