Multimodal Reasoning on GEO-AT RNA
0.491METEORCuttlefish + Qwen3-8B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Cuttlefish + Qwen3-8BArchitecture=Cuttlefish augmentation, Backbone Model=Qwen3-8B2026.02 | 0.491 | 0.89 | |
| Cuttlefish + R1-7BArchitecture=Cuttlefish augmentation, Backbone Model=DeepSeek-R1-D-Qwen-7B2026.02 | 0.422 | 0.799 | |
| Cuttlefish + Llama-3.1-8B-IArchitecture=Cuttlefish augmentation, Backbone Model=Llama-3.1-8B-Instruct2026.02 | 0.403 | 0.868 | |
| Cuttlefish + GLM-4-9BArchitecture=Cuttlefish augmentation, Backbone Model=GLM-4-9B2026.02 | 0.403 | 0.794 | |
| Cuttlefish + Mistral3-8B-RArchitecture=Cuttlefish augmentation, Backbone Model=Mistral3-8B-Reasoning2026.02 | 0.395 | 0.775 | |
| Cuttlefish + Qwen2.5-7BArchitecture=Cuttlefish augmentation, Backbone Model=Qwen2.5-7B2026.02 | 0.378 | 0.858 | |
| Mistral-3-8B-Reasoning-2512Reasoning Capability=Reasoning, Modality Handling=Modality Enhanced Tokenizer2026.02 | 0.288 | 0.579 | |
| DeepSeek-R1-D-Qwen-7BReasoning Capability=Reasoning, Modality Handling=Modality Enhanced Tokenizer2026.02 | 0.283 | 0.688 | |
| Cuttlefish + Mistral-3-8B-IArchitecture=Cuttlefish augmentation, Backbone Model=Mistral-3-8B-Instruct2026.02 | 0.22 | 0.694 | |
| Mistral-3-8B-Instruct-2512Reasoning Capability=Non-Reasoning, Modality Handling=Sequence Only2026.02 | 0.192 | 0.652 | |
| Mistral-3-8B-Reasoning-2512Reasoning Capability=Reasoning, Modality Handling=Sequence Only2026.02 | 0.18 | 0.674 | |
| Llama-3.1-8B-InstructReasoning Capability=Non-Reasoning, Modality Handling=Sequence Only2026.02 | 0.175 | 0.646 | |
| DeepSeek-R1-D-Qwen-7BReasoning Capability=Reasoning, Modality Handling=Sequence Only2026.02 | 0.17 | 0.687 | |
| Qwen-2.5-7B-InstructReasoning Capability=Non-Reasoning, Modality Handling=Sequence Only2026.02 | 0.159 | 0.64 | |
| GLM-4-9B-0414Reasoning Capability=Non-Reasoning, Modality Handling=Sequence Only2026.02 | 0.143 | 0.549 | |
| Qwen3-8BReasoning Capability=Reasoning, Modality Handling=Sequence Only2026.02 | 0.128 | 0.655 | |
| Qwen3-8BReasoning Capability=Reasoning, Modality Handling=Modality Enhanced Tokenizer2026.02 | 0.095 | 0.773 |