# Open-ended Writing on ArenaHard Source: [Open-ended Writing on ArenaHard]() Collection index: [llms.txt]() Data updated: 2026-06-02T10:46:50.366Z Evaluation of model performance on the ArenaHard dataset for Open-ended Writing tasks, measured by accuracy. 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) | ArenaHard | | Metric | Accuracy | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 17 of 17 | | 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: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Official Instruct Model | 50 | Not recorded | 6a1e640adbd7150506f8564a | | Opus-4.7 (xHigh) | 33.84 | Not recorded | 6a1e640adbd7150506f8564a | | GLM-4.7 & ANDES | 12.94 | Not recorded | 6a1e640adbd7150506f8564a | | GLM-4.7 (Scaffold-only) | 5.7 | Not recorded | 6a1e640adbd7150506f8564a | | Opus-4.6 (1M) | 3.21 | Not recorded | 6a1e640adbd7150506f8564a | | Gemini-3.1-Pro | 2.27 | Not recorded | 6a1e640adbd7150506f8564a | | MiniMax-M2.5 | 1.76 | Not recorded | 6a1e640adbd7150506f8564a | | GPT-5.2 | 1.27 | Not recorded | 6a1e640adbd7150506f8564a | | Opus-4.6 | 1.15 | Not recorded | 6a1e640adbd7150506f8564a | | Base Model (Qwen3-1.7B) | 0.91 | Not recorded | 6a1e640adbd7150506f8564a | | MiniMax-M2.1 | 0.91 | Not recorded | 6a1e640adbd7150506f8564a | | Qwen3-Max | 0.91 | Not recorded | 6a1e640adbd7150506f8564a | | Kimi-K2-Thinking | 0.91 | Not recorded | 6a1e640adbd7150506f8564a | | GPT-5.4 (High) | 0.91 | Not recorded | 6a1e640adbd7150506f8564a | | GLM-4.7 (OpenCode) | 0.91 | Not recorded | 6a1e640adbd7150506f8564a | | Sonnet-4.5 | 0.21 | Not recorded | 6a1e640adbd7150506f8564a | | GPT-5.1-Codex-Max | 0.14 | Not recorded | 6a1e640adbd7150506f8564a | ### Official Instruct Model Accuracy: 50. Result ID: 6a1eb47cfbc373592c19a1e7. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Official Instruct Model"}. This record does not assert independent reproduction. ### Opus-4.7 (xHigh) Accuracy: 33.84. Result ID: 6a1eb47cfbc373592c19a1f4. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### GLM-4.7 & ANDES Accuracy: 12.94. Result ID: 6a1eb47cfbc373592c19a1f7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Category":"Proposed Method Comparison","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### GLM-4.7 (Scaffold-only) Accuracy: 5.7. Result ID: 6a1eb47cfbc373592c19a1f6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Category":"Proposed Method Comparison","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Opus-4.6 (1M) Accuracy: 3.21. Result ID: 6a1eb47cfbc373592c19a1f3. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Gemini-3.1-Pro Accuracy: 2.27. Result ID: 6a1eb47cfbc373592c19a1f1. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### MiniMax-M2.5 Accuracy: 1.76. Result ID: 6a1eb47cfbc373592c19a1ea. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### GPT-5.2 Accuracy: 1.27. Result ID: 6a1eb47cfbc373592c19a1f0. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Opus-4.6 Accuracy: 1.15. Result ID: 6a1eb47cfbc373592c19a1ef. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Base Model (Qwen3-1.7B) Accuracy: 0.91. Result ID: 6a1eb47cfbc373592c19a1e8. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Zero-Shot"}. This record does not assert independent reproduction. ### MiniMax-M2.1 Accuracy: 0.91. Result ID: 6a1eb47cfbc373592c19a1eb. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Qwen3-Max Accuracy: 0.91. Result ID: 6a1eb47cfbc373592c19a1ec. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Kimi-K2-Thinking Accuracy: 0.91. Result ID: 6a1eb47cfbc373592c19a1ed. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### GPT-5.4 (High) Accuracy: 0.91. Result ID: 6a1eb47cfbc373592c19a1f2. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### GLM-4.7 (OpenCode) Accuracy: 0.91. Result ID: 6a1eb47cfbc373592c19a1f5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Category":"Proposed Method Comparison","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Sonnet-4.5 Accuracy: 0.21. Result ID: 6a1eb47cfbc373592c19a1e9. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### GPT-5.1-Codex-Max Accuracy: 0.14. Result ID: 6a1eb47cfbc373592c19a1ee. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Category":"Agent-Post-Trained","Base Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.