{ "leaderboard_info": { "total_models": 18 }, "models": [ { "name": "Claude-3.7-Sonnet", "name_link": "https://www.anthropic.com/claude", "submitter": "Anthropic Team", "submitter_link": "mailto:support@anthropic.com", "submission_time": "2025-08-01T17:09:29.917540Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 75.78, "subcategories": { "Spectrum Type Classification": { "accuracy": 96.36 }, "Spectrum Quality Assessment": { "accuracy": 38.33 }, "Basic Feature Extraction": { "accuracy": 86.27 }, "Impurity Peak Detection": { "accuracy": 82.14 } } }, "Perception": { "accuracy": 79.9, "subcategories": { "Functional Group Recognition": { "accuracy": 71.43 }, "Elemental Compositional Prediction": { "accuracy": 88.89 }, "Peak Assignment": { "accuracy": 71.05 }, "Basic Property Prediction": { "accuracy": 88.24 } } }, "Semantic": { "accuracy": 81.94, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 82.28 }, "Fusing Spectroscopic Modalities": { "accuracy": 74.36 }, "Multimodal Molecular Reasoning": { "accuracy": 89.19 } } }, "Generation": { "accuracy": 8.42, "subcategories": { "Forward Problems": { "accuracy": 20.0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 5.26 } } }, "overall_accuracy": 61.51 }, "model_info": { "homepage": "https://www.anthropic.com/claude", "paper": "", "code": "", "description": "Claude 3.7 Sonnet with enhanced multimodal capabilities" } }, { "name": "Doubao-1.5-Vision-Pro-Thinking", "name_link": "https://www.volcengine.com/product/doubao", "submitter": "ByteDance Team", "submitter_link": "https://www.volcengine.com/", "submission_time": "2025-08-01T17:09:29.934647Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 69.41, "subcategories": { "Spectrum Type Classification": { "accuracy": 96.36 }, "Spectrum Quality Assessment": { "accuracy": 35.0 }, "Basic Feature Extraction": { "accuracy": 78.43 }, "Impurity Peak Detection": { "accuracy": 67.86 } } }, "Perception": { "accuracy": 74.69, "subcategories": { "Functional Group Recognition": { "accuracy": 53.33 }, "Elemental Compositional Prediction": { "accuracy": 80.56 }, "Peak Assignment": { "accuracy": 73.68 }, "Basic Property Prediction": { "accuracy": 91.18 } } }, "Semantic": { "accuracy": 75.77, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 68.75 }, "Fusing Spectroscopic Modalities": { "accuracy": 66.67 }, "Multimodal Molecular Reasoning": { "accuracy": 91.89 } } }, "Generation": { "accuracy": 25.64, "subcategories": { "Forward Problems": { "accuracy": 66.67 }, "Inverse Problems": { "accuracy": 5.0 }, "De Novo Generation": { "accuracy": 5.26 } } }, "overall_accuracy": 61.38 }, "model_info": { "homepage": "https://www.volcengine.com/product/doubao", "paper": "", "code": "", "description": "Doubao 1.5 Vision Pro with enhanced reasoning capabilities" } }, { "name": "Qwen2.5-VL-72B-Instruct", "name_link": "https://qwenlm.github.io/", "submitter": "Alibaba DAMO Academy", "submitter_link": "https://damo.alibaba.com/", "submission_time": "2025-08-01T17:09:29.925741Z", "model_type": "open_source", "model_size": "72B", "is_multimodal": true, "results": { "Signal": { "accuracy": 78.0, "subcategories": { "Spectrum Type Classification": { "accuracy": 94.55 }, "Spectrum Quality Assessment": { "accuracy": 38.33 }, "Basic Feature Extraction": { "accuracy": 86.27 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 72.49, "subcategories": { "Functional Group Recognition": { "accuracy": 42.22 }, "Elemental Compositional Prediction": { "accuracy": 80.56 }, "Peak Assignment": { "accuracy": 78.95 }, "Basic Property Prediction": { "accuracy": 88.24 } } }, "Semantic": { "accuracy": 78.35, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 66.25 }, "Fusing Spectroscopic Modalities": { "accuracy": 76.92 }, "Multimodal Molecular Reasoning": { "accuracy": 91.89 } } }, "Generation": { "accuracy": 13.51, "subcategories": { "Forward Problems": { "accuracy": 30.0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 10.53 } } }, "overall_accuracy": 60.59 }, "model_info": { "homepage": "https://qwenlm.github.io/", "paper": "https://arxiv.org/abs/2407.10671", "code": "https://github.com/QwenLM/Qwen2.5", "description": "Qwen2.5-VL-72B large-scale open-source vision-language model" } }, { "name": "GPT-4.1", "name_link": "https://openai.com/gpt-4", "submitter": "OpenAI Team", "submitter_link": "mailto:research@openai.com", "submission_time": "2025-08-01T17:09:29.920918Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 73.72, "subcategories": { "Spectrum Type Classification": { "accuracy": 94.55 }, "Spectrum Quality Assessment": { "accuracy": 28.33 }, "Basic Feature Extraction": { "accuracy": 86.27 }, "Impurity Peak Detection": { "accuracy": 85.71 } } }, "Perception": { "accuracy": 68.42, "subcategories": { "Functional Group Recognition": { "accuracy": 53.33 }, "Elemental Compositional Prediction": { "accuracy": 77.78 }, "Peak Assignment": { "accuracy": 63.16 }, "Basic Property Prediction": { "accuracy": 79.41 } } }, "Semantic": { "accuracy": 80.35, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 82.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 66.67 }, "Multimodal Molecular Reasoning": { "accuracy": 91.89 } } }, "Generation": { "accuracy": 14.62, "subcategories": { "Forward Problems": { "accuracy": 33.33 }, "Inverse Problems": { "accuracy": 10.53 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 59.28 }, "model_info": { "homepage": "https://openai.com/gpt-4", "paper": "https://arxiv.org/abs/2303.08774", "code": "", "description": "GPT-4.1 with improved reasoning capabilities" } }, { "name": "Qwen-VL-Max", "name_link": "https://qwenlm.github.io/", "submitter": "Alibaba DAMO Academy", "submitter_link": "https://damo.alibaba.com/", "submission_time": "2025-08-01T17:09:29.923747Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 78.57, "subcategories": { "Spectrum Type Classification": { "accuracy": 94.55 }, "Spectrum Quality Assessment": { "accuracy": 36.67 }, "Basic Feature Extraction": { "accuracy": 90.2 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 76.94, "subcategories": { "Functional Group Recognition": { "accuracy": 60.0 }, "Elemental Compositional Prediction": { "accuracy": 80.56 }, "Peak Assignment": { "accuracy": 78.95 }, "Basic Property Prediction": { "accuracy": 88.24 } } }, "Semantic": { "accuracy": 65.39, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 32.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 71.79 }, "Multimodal Molecular Reasoning": { "accuracy": 91.89 } } }, "Generation": { "accuracy": 16.2, "subcategories": { "Forward Problems": { "accuracy": 43.33 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 5.26 } } }, "overall_accuracy": 59.27 }, "model_info": { "homepage": "https://qwenlm.github.io/", "paper": "https://arxiv.org/abs/2308.12966", "code": "https://github.com/QwenLM/Qwen-VL", "description": "Qwen-VL-Max with advanced vision-language understanding" } }, { "name": "GPT-4-Vision", "name_link": "https://openai.com/gpt-4", "submitter": "OpenAI Team", "submitter_link": "mailto:research@openai.com", "submission_time": "2025-08-01T17:09:29.921628Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 73.32, "subcategories": { "Spectrum Type Classification": { "accuracy": 94.55 }, "Spectrum Quality Assessment": { "accuracy": 33.33 }, "Basic Feature Extraction": { "accuracy": 72.55 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 74.74, "subcategories": { "Functional Group Recognition": { "accuracy": 73.33 }, "Elemental Compositional Prediction": { "accuracy": 72.22 }, "Peak Assignment": { "accuracy": 71.05 }, "Basic Property Prediction": { "accuracy": 82.35 } } }, "Semantic": { "accuracy": 74.97, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 73.75 }, "Fusing Spectroscopic Modalities": { "accuracy": 53.85 }, "Multimodal Molecular Reasoning": { "accuracy": 97.3 } } }, "Generation": { "accuracy": 9.44, "subcategories": { "Forward Problems": { "accuracy": 23.33 }, "Inverse Problems": { "accuracy": 5.0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 58.12 }, "model_info": { "homepage": "https://openai.com/gpt-4", "paper": "https://arxiv.org/abs/2303.08774", "code": "", "description": "GPT-4 with vision capabilities for multimodal understanding" } }, { "name": "Claude-3.5-Sonnet", "name_link": "https://www.anthropic.com/claude", "submitter": "Anthropic Team", "submitter_link": "mailto:support@anthropic.com", "submission_time": "2025-08-01T17:09:29.916821Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 68.15, "subcategories": { "Spectrum Type Classification": { "accuracy": 96.36 }, "Spectrum Quality Assessment": { "accuracy": 28.33 }, "Basic Feature Extraction": { "accuracy": 76.47 }, "Impurity Peak Detection": { "accuracy": 71.43 } } }, "Perception": { "accuracy": 74.85, "subcategories": { "Functional Group Recognition": { "accuracy": 60.0 }, "Elemental Compositional Prediction": { "accuracy": 77.78 }, "Peak Assignment": { "accuracy": 76.32 }, "Basic Property Prediction": { "accuracy": 85.29 } } }, "Semantic": { "accuracy": 82.11, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 82.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 69.23 }, "Multimodal Molecular Reasoning": { "accuracy": 94.59 } } }, "Generation": { "accuracy": 6.67, "subcategories": { "Forward Problems": { "accuracy": 20.0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 57.95 }, "model_info": { "homepage": "https://www.anthropic.com/claude", "paper": "", "code": "", "description": "Claude 3.5 Sonnet with advanced reasoning and multimodal capabilities" } }, { "name": "Claude-4-Opus", "name_link": "https://www.anthropic.com/claude", "submitter": "Anthropic Team", "submitter_link": "mailto:support@anthropic.com", "submission_time": "2025-08-01T17:09:29.919719Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 77.2, "subcategories": { "Spectrum Type Classification": { "accuracy": 96.36 }, "Spectrum Quality Assessment": { "accuracy": 33.33 }, "Basic Feature Extraction": { "accuracy": 86.27 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 78.25, "subcategories": { "Functional Group Recognition": { "accuracy": 73.33 }, "Elemental Compositional Prediction": { "accuracy": 83.33 }, "Peak Assignment": { "accuracy": 71.05 }, "Basic Property Prediction": { "accuracy": 85.29 } } }, "Semantic": { "accuracy": 65.3, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 32.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 76.92 }, "Multimodal Molecular Reasoning": { "accuracy": 86.49 } } }, "Generation": { "accuracy": 7.31, "subcategories": { "Forward Problems": { "accuracy": 16.67 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 5.26 } } }, "overall_accuracy": 57.02 }, "model_info": { "homepage": "https://www.anthropic.com/claude", "paper": "", "code": "", "description": "Claude 4 Opus - most capable model in the Claude family" } }, { "name": "Doubao-1.5-Vision-Pro", "name_link": "https://www.volcengine.com/product/doubao", "submitter": "ByteDance Team", "submitter_link": "https://www.volcengine.com/", "submission_time": "2025-08-01T17:09:29.932862Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 75.7, "subcategories": { "Spectrum Type Classification": { "accuracy": 98.18 }, "Spectrum Quality Assessment": { "accuracy": 33.33 }, "Basic Feature Extraction": { "accuracy": 78.43 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 76.66, "subcategories": { "Functional Group Recognition": { "accuracy": 66.67 }, "Elemental Compositional Prediction": { "accuracy": 83.33 }, "Peak Assignment": { "accuracy": 68.42 }, "Basic Property Prediction": { "accuracy": 88.24 } } }, "Semantic": { "accuracy": 71.03, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 67.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 56.41 }, "Multimodal Molecular Reasoning": { "accuracy": 89.19 } } }, "Generation": { "accuracy": 2.22, "subcategories": { "Forward Problems": { "accuracy": 6.67 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 56.4 }, "model_info": { "homepage": "https://www.volcengine.com/product/doubao", "paper": "", "code": "", "description": "Doubao 1.5 Vision Pro multimodal model" } }, { "name": "InternVL3-78B", "name_link": "https://internvl.github.io/", "submitter": "Shanghai AI Laboratory", "submitter_link": "https://www.shlab.org.cn/", "submission_time": "2025-08-01T17:09:29.926853Z", "model_type": "open_source", "model_size": "78B", "is_multimodal": true, "results": { "Signal": { "accuracy": 69.18, "subcategories": { "Spectrum Type Classification": { "accuracy": 96.36 }, "Spectrum Quality Assessment": { "accuracy": 38.33 }, "Basic Feature Extraction": { "accuracy": 70.59 }, "Impurity Peak Detection": { "accuracy": 71.43 } } }, "Perception": { "accuracy": 73.33, "subcategories": { "Functional Group Recognition": { "accuracy": 48.49 }, "Elemental Compositional Prediction": { "accuracy": 75.0 }, "Peak Assignment": { "accuracy": 81.58 }, "Basic Property Prediction": { "accuracy": 88.24 } } }, "Semantic": { "accuracy": 71.84, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 62.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 69.23 }, "Multimodal Molecular Reasoning": { "accuracy": 83.78 } } }, "Generation": { "accuracy": 9.53, "subcategories": { "Forward Problems": { "accuracy": 23.33 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 5.26 } } }, "overall_accuracy": 55.97 }, "model_info": { "homepage": "https://internvl.github.io/", "paper": "https://arxiv.org/abs/2312.14238", "code": "https://github.com/OpenGVLab/InternVL", "description": "InternVL3-78B large-scale multimodal foundation model" } }, { "name": "GPT-4o", "name_link": "https://openai.com/gpt-4", "submitter": "OpenAI Team", "submitter_link": "mailto:research@openai.com", "submission_time": "2025-08-01T17:09:29.920289Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 72.8, "subcategories": { "Spectrum Type Classification": { "accuracy": 96.36 }, "Spectrum Quality Assessment": { "accuracy": 33.33 }, "Basic Feature Extraction": { "accuracy": 68.63 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 69.53, "subcategories": { "Functional Group Recognition": { "accuracy": 57.78 }, "Elemental Compositional Prediction": { "accuracy": 77.78 }, "Peak Assignment": { "accuracy": 63.16 }, "Basic Property Prediction": { "accuracy": 79.41 } } }, "Semantic": { "accuracy": 77.44, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 78.75 }, "Fusing Spectroscopic Modalities": { "accuracy": 58.97 }, "Multimodal Molecular Reasoning": { "accuracy": 94.59 } } }, "Generation": { "accuracy": 3.33, "subcategories": { "Forward Problems": { "accuracy": 10.0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 55.77 }, "model_info": { "homepage": "https://openai.com/gpt-4", "paper": "https://arxiv.org/abs/2303.08774", "code": "", "description": "GPT-4o with enhanced multimodal capabilities" } }, { "name": "Grok-2-Vision", "name_link": "https://grok.x.ai/", "submitter": "xAI Team", "submitter_link": "mailto:team@x.ai", "submission_time": "2025-08-01T17:09:29.922697Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 72.5, "subcategories": { "Spectrum Type Classification": { "accuracy": 94.55 }, "Spectrum Quality Assessment": { "accuracy": 31.67 }, "Basic Feature Extraction": { "accuracy": 74.51 }, "Impurity Peak Detection": { "accuracy": 89.29 } } }, "Perception": { "accuracy": 75.26, "subcategories": { "Functional Group Recognition": { "accuracy": 64.44 }, "Elemental Compositional Prediction": { "accuracy": 80.56 }, "Peak Assignment": { "accuracy": 73.68 }, "Basic Property Prediction": { "accuracy": 82.35 } } }, "Semantic": { "accuracy": 61.75, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 37.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 66.67 }, "Multimodal Molecular Reasoning": { "accuracy": 81.08 } } }, "Generation": { "accuracy": 7.78, "subcategories": { "Forward Problems": { "accuracy": 23.33 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 54.32 }, "model_info": { "homepage": "https://grok.x.ai/", "paper": "", "code": "", "description": "Grok-2 with vision capabilities for multimodal reasoning" } }, { "name": "Claude-3.5-Haiku", "name_link": "https://www.anthropic.com/claude", "submitter": "Anthropic Team", "submitter_link": "mailto:support@anthropic.com", "submission_time": "2025-08-01T17:09:29.919136Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 67.52, "subcategories": { "Spectrum Type Classification": { "accuracy": 94.55 }, "Spectrum Quality Assessment": { "accuracy": 31.67 }, "Basic Feature Extraction": { "accuracy": 50.98 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 73.61, "subcategories": { "Functional Group Recognition": { "accuracy": 66.67 }, "Elemental Compositional Prediction": { "accuracy": 75.0 }, "Peak Assignment": { "accuracy": 76.32 }, "Basic Property Prediction": { "accuracy": 76.47 } } }, "Semantic": { "accuracy": 70.89, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 67.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 64.1 }, "Multimodal Molecular Reasoning": { "accuracy": 81.08 } } }, "Generation": { "accuracy": 3.33, "subcategories": { "Forward Problems": { "accuracy": 10.0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 53.84 }, "model_info": { "homepage": "https://www.anthropic.com/claude", "paper": "", "code": "", "description": "Claude 3.5 Haiku - fast and efficient multimodal model" } }, { "name": "Claude-4-Sonnet", "name_link": "https://www.anthropic.com/claude", "submitter": "Anthropic Team", "submitter_link": "mailto:support@anthropic.com", "submission_time": "2025-08-01T17:09:29.918518Z", "model_type": "proprietary", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 78.11, "subcategories": { "Spectrum Type Classification": { "accuracy": 96.36 }, "Spectrum Quality Assessment": { "accuracy": 35.0 }, "Basic Feature Extraction": { "accuracy": 88.24 }, "Impurity Peak Detection": { "accuracy": 92.86 } } }, "Perception": { "accuracy": 65.78, "subcategories": { "Functional Group Recognition": { "accuracy": 62.22 }, "Elemental Compositional Prediction": { "accuracy": 63.89 }, "Peak Assignment": { "accuracy": 60.53 }, "Basic Property Prediction": { "accuracy": 76.47 } } }, "Semantic": { "accuracy": 41.57, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 16.25 }, "Fusing Spectroscopic Modalities": { "accuracy": 43.59 }, "Multimodal Molecular Reasoning": { "accuracy": 64.86 } } }, "Generation": { "accuracy": 8.13, "subcategories": { "Forward Problems": { "accuracy": 3.33 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 21.05 } } }, "overall_accuracy": 48.4 }, "model_info": { "homepage": "https://www.anthropic.com/claude", "paper": "", "code": "", "description": "Claude 4 Sonnet with advanced reasoning capabilities" } }, { "name": "Qwen2.5-VL-32B-Instruct", "name_link": "https://qwenlm.github.io/", "submitter": "Alibaba DAMO Academy", "submitter_link": "https://damo.alibaba.com/", "submission_time": "2025-08-01T17:09:29.924718Z", "model_type": "open_source", "model_size": "32B", "is_multimodal": true, "results": { "Signal": { "accuracy": 57.02, "subcategories": { "Spectrum Type Classification": { "accuracy": 92.73 }, "Spectrum Quality Assessment": { "accuracy": 26.67 }, "Basic Feature Extraction": { "accuracy": 37.25 }, "Impurity Peak Detection": { "accuracy": 71.43 } } }, "Perception": { "accuracy": 48.89, "subcategories": { "Functional Group Recognition": { "accuracy": 57.78 }, "Elemental Compositional Prediction": { "accuracy": 44.44 }, "Peak Assignment": { "accuracy": 31.58 }, "Basic Property Prediction": { "accuracy": 61.76 } } }, "Semantic": { "accuracy": 17.03, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 0.0 }, "Fusing Spectroscopic Modalities": { "accuracy": 5.13 }, "Multimodal Molecular Reasoning": { "accuracy": 45.95 } } }, "Generation": { "accuracy": 6.67, "subcategories": { "Forward Problems": { "accuracy": 20.0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 32.4 }, "model_info": { "homepage": "https://qwenlm.github.io/", "paper": "https://arxiv.org/abs/2407.10671", "code": "https://github.com/QwenLM/Qwen2.5", "description": "Qwen2.5-VL-32B open-source vision-language model" } }, { "name": "DeepSeek-VL2", "name_link": "https://www.deepseek.com/", "submitter": "DeepSeek Team", "submitter_link": "https://github.com/deepseek-ai", "submission_time": "2025-08-01T17:09:29.931327Z", "model_type": "open_source", "model_size": "Unknown", "is_multimodal": true, "results": { "Signal": { "accuracy": 33.51, "subcategories": { "Spectrum Type Classification": { "accuracy": 52.73 }, "Spectrum Quality Assessment": { "accuracy": 23.33 }, "Basic Feature Extraction": { "accuracy": 29.41 }, "Impurity Peak Detection": { "accuracy": 28.57 } } }, "Perception": { "accuracy": 28.91, "subcategories": { "Functional Group Recognition": { "accuracy": 8.89 }, "Elemental Compositional Prediction": { "accuracy": 27.78 }, "Peak Assignment": { "accuracy": 28.95 }, "Basic Property Prediction": { "accuracy": 50.0 } } }, "Semantic": { "accuracy": 20.94, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 15.0 }, "Fusing Spectroscopic Modalities": { "accuracy": 15.38 }, "Multimodal Molecular Reasoning": { "accuracy": 32.43 } } }, "Generation": { "accuracy": 6.75, "subcategories": { "Forward Problems": { "accuracy": 10.0 }, "Inverse Problems": { "accuracy": 5.0 }, "De Novo Generation": { "accuracy": 5.26 } } }, "overall_accuracy": 22.53 }, "model_info": { "homepage": "https://www.deepseek.com/", "paper": "https://arxiv.org/abs/2412.10302", "code": "https://github.com/deepseek-ai/DeepSeek-VL2", "description": "DeepSeek-VL2 advanced vision-language model" } }, { "name": "Llama-3.2-90B-Vision-Instruct", "name_link": "https://llama.meta.com/", "submitter": "Meta AI", "submitter_link": "https://ai.meta.com/", "submission_time": "2025-08-01T17:09:29.929751Z", "model_type": "open_source", "model_size": "90B", "is_multimodal": true, "results": { "Signal": { "accuracy": 27.12, "subcategories": { "Spectrum Type Classification": { "accuracy": 38.18 }, "Spectrum Quality Assessment": { "accuracy": 10.0 }, "Basic Feature Extraction": { "accuracy": 35.29 }, "Impurity Peak Detection": { "accuracy": 25.0 } } }, "Perception": { "accuracy": 23.77, "subcategories": { "Functional Group Recognition": { "accuracy": 17.78 }, "Elemental Compositional Prediction": { "accuracy": 27.78 }, "Peak Assignment": { "accuracy": 28.95 }, "Basic Property Prediction": { "accuracy": 20.59 } } }, "Semantic": { "accuracy": 23.21, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 21.25 }, "Fusing Spectroscopic Modalities": { "accuracy": 5.13 }, "Multimodal Molecular Reasoning": { "accuracy": 43.24 } } }, "Generation": { "accuracy": 0, "subcategories": { "Forward Problems": { "accuracy": 0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 18.52 }, "model_info": { "homepage": "https://llama.meta.com/", "paper": "https://arxiv.org/abs/2407.21783", "code": "https://github.com/meta-llama/llama3", "description": "Llama 3.2 90B with vision capabilities" } }, { "name": "Llama-3.2-11B-Vision-Instruct", "name_link": "https://llama.meta.com/", "submitter": "Meta AI", "submitter_link": "https://ai.meta.com/", "submission_time": "2025-08-01T17:09:29.928044Z", "model_type": "open_source", "model_size": "11B", "is_multimodal": true, "results": { "Signal": { "accuracy": 21.24, "subcategories": { "Spectrum Type Classification": { "accuracy": 34.55 }, "Spectrum Quality Assessment": { "accuracy": 11.67 }, "Basic Feature Extraction": { "accuracy": 13.73 }, "Impurity Peak Detection": { "accuracy": 25.0 } } }, "Perception": { "accuracy": 26.72, "subcategories": { "Functional Group Recognition": { "accuracy": 20.0 }, "Elemental Compositional Prediction": { "accuracy": 41.67 }, "Peak Assignment": { "accuracy": 15.79 }, "Basic Property Prediction": { "accuracy": 29.41 } } }, "Semantic": { "accuracy": 11.42, "subcategories": { "Molecular Structure Elucidation": { "accuracy": 7.5 }, "Fusing Spectroscopic Modalities": { "accuracy": 5.13 }, "Multimodal Molecular Reasoning": { "accuracy": 21.62 } } }, "Generation": { "accuracy": 0, "subcategories": { "Forward Problems": { "accuracy": 0 }, "Inverse Problems": { "accuracy": 0 }, "De Novo Generation": { "accuracy": 0 } } }, "overall_accuracy": 14.84 }, "model_info": { "homepage": "https://llama.meta.com/", "paper": "https://arxiv.org/abs/2407.21783", "code": "https://github.com/meta-llama/llama3", "description": "Llama 3.2 11B with vision capabilities" } } ] }