# Open-ended writing on DeepresearchBench Source: [Open-ended writing on DeepresearchBench]() Collection index: [llms.txt]() Data updated: 2026-04-06T08:44:15.082Z Evaluation of model performance on open-ended writing tasks using the DeepresearchBench dataset, measured by Overall, Comprehensiveness (Comp.), Depth, Instruction adherence (Inst.), and Readability (Read) 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 | Open-ended writing | | Dataset (including any stated split) | DeepresearchBench | | Metric | Overall Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 11 of 11 | | 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 - [Overall Score](): Higher is better; unit: PERCENT. - [Comprehensiveness](): Higher is better; unit: PERCENT. - [Depth](): Higher is better; unit: PERCENT. - [Instruction Adherence](): Higher is better; unit: PERCENT. - [Readability](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Score | Result date | Report paper | | --- | --- | --- | --- | | Qwen3-8B + R2-Write-SFT + RLp | 46.93 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + R2-Write-SFT | 46 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B + RL | 44.97 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B + R2-Write-SFT + RLp | 44.95 | Not recorded | 69d33ded7d0172140ec735d6 | | LongWriter-Zero-32B | 44.69 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B + R2-Write-SFT | 44.18 | Not recorded | 69d33ded7d0172140ec735d6 | | Reverse-Engineering-8B | 42.42 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-8B | 40.52 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B + RL | 39.74 | Not recorded | 69d33ded7d0172140ec735d6 | | Longwriter-9B | 38.68 | Not recorded | 69d33ded7d0172140ec735d6 | | Qwen3-4B | 38.11 | Not recorded | 69d33ded7d0172140ec735d6 | ### Qwen3-8B + R2-Write-SFT + RLp Overall Score: 46.93. Result ID: 69d3725e47f34ce924bb5c6d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + PPO"}. This record does not assert independent reproduction. ### Qwen3-8B + R2-Write-SFT Overall Score: 46. Result ID: 69d3725e47f34ce924bb5c68. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-8B + RL Overall Score: 44.97. Result ID: 69d3725e47f34ce924bb5c6a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"PPO"}. This record does not assert independent reproduction. ### Qwen3-4B + R2-Write-SFT + RLp Overall Score: 44.95. Result ID: 69d3725e47f34ce924bb5c6c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + PPO"}. This record does not assert independent reproduction. ### LongWriter-Zero-32B Overall Score: 44.69. Result ID: 69d3725e47f34ce924bb5c6b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT + GRPO"}. This record does not assert independent reproduction. ### Qwen3-4B + R2-Write-SFT Overall Score: 44.18. Result ID: 69d3725e47f34ce924bb5c67. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Reverse-Engineering-8B Overall Score: 42.42. Result ID: 69d3725e47f34ce924bb5c66. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-8B Overall Score: 40.52. Result ID: 69d3725e47f34ce924bb5c64. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"-"}. This record does not assert independent reproduction. ### Qwen3-4B + RL Overall Score: 39.74. Result ID: 69d3725e47f34ce924bb5c69. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"PPO"}. This record does not assert independent reproduction. ### Longwriter-9B Overall Score: 38.68. Result ID: 69d3725e47f34ce924bb5c65. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training":"SFT"}. This record does not assert independent reproduction. ### Qwen3-4B Overall Score: 38.11. Result ID: 69d3725e47f34ce924bb5c63. [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.