# Reasoning-based Text-to-Image Generation on T2I-CoREBench Source: [Reasoning-based Text-to-Image Generation on T2I-CoREBench]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:59:14.085Z Evaluation of Text-to-Image models on the T2I-CoREBench dataset, focusing on reasoning capabilities across various sub-dimensions (e.g., LR, BR, HR, PR). 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 | Reasoning-based Text-to-Image Generation | | Dataset (including any stated split) | T2I-CoREBench | | Metric | R-LR Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 15 of 15 | | 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 - [R-LR Score](): Higher is better; unit: PERCENT. - [R-BR Score](): Higher is better; unit: PERCENT. - [R-HR Score](): Higher is better; unit: PERCENT. - [R-PR Score](): Higher is better; unit: PERCENT. - [R-GR Score](): Higher is better; unit: PERCENT. - [R-AR Score](): Higher is better; unit: PERCENT. - [R-CR Score](): Higher is better; unit: PERCENT. - [R-RR Score](): Higher is better; unit: PERCENT. - [Overall Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | R-LR Score | Result date | Report paper | | --- | --- | --- | --- | | Seedream 4.0 | 79.2 | Not recorded | 698eb058c97a0b3e3b5e695b | | Nano Banana | 65.4 | Not recorded | 698eb058c97a0b3e3b5e695b | | GPT-Image-1 | 61.6 | Not recorded | 698eb058c97a0b3e3b5e695b | | Qwen-Image | 42.2 | Not recorded | 698eb058c97a0b3e3b5e695b | | LongCat-Image | 41.7 | Not recorded | 698eb058c97a0b3e3b5e695b | | Hunyuan-Image 3.0 | 41.6 | Not recorded | 698eb058c97a0b3e3b5e695b | | DeepGen 1.0 (SFT) | 38.8 | Not recorded | 698eb058c97a0b3e3b5e695b | | DeepGen 1.0 (RL) | 38.5 | Not recorded | 698eb058c97a0b3e3b5e695b | | Z-Image-Turbo | 37.8 | Not recorded | 698eb058c97a0b3e3b5e695b | | Show-o2 | 30.2 | Not recorded | 698eb058c97a0b3e3b5e695b | | BAGEL | 28.6 | Not recorded | 698eb058c97a0b3e3b5e695b | | Janus-Pro | 27.2 | Not recorded | 698eb058c97a0b3e3b5e695b | | OmniGen2 | 26.8 | Not recorded | 698eb058c97a0b3e3b5e695b | | FLUX.1 \[Dev\] | 26.3 | Not recorded | 698eb058c97a0b3e3b5e695b | | BLIP3-o | 18.4 | Not recorded | 698eb058c97a0b3e3b5e695b | ### Seedream 4.0 R-LR Score: 79.2. Result ID: 69970c705698665ec459e291. [Reporting paper]() Recorded evaluation configuration: {"Access":"Closed-source"}. This record does not assert independent reproduction. ### Nano Banana R-LR Score: 65.4. Result ID: 69970c705698665ec459e28f. [Reporting paper]() Recorded evaluation configuration: {"Access":"Closed-source"}. This record does not assert independent reproduction. ### GPT-Image-1 R-LR Score: 61.6. Result ID: 69970c705698665ec459e290. [Reporting paper]() Recorded evaluation configuration: {"Access":"Closed-source"}. This record does not assert independent reproduction. ### Qwen-Image R-LR Score: 42.2. Result ID: 69970c705698665ec459e299. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-source","Params":"7B + 20B"}. This record does not assert independent reproduction. ### LongCat-Image R-LR Score: 41.7. Result ID: 69970c705698665ec459e29b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-source","Params":"7B + 6B"}. This record does not assert independent reproduction. ### Hunyuan-Image 3.0 R-LR Score: 41.6. Result ID: 69970c705698665ec459e298. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-source","Params":"80B"}. This record does not assert independent reproduction. ### DeepGen 1.0 (SFT) R-LR Score: 38.8. Result ID: 69970c705698665ec459e29c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-source","Params":"3B + 2B","Mode":"SFT"}. This record does not assert independent reproduction. ### DeepGen 1.0 (RL) R-LR Score: 38.5. Result ID: 69970c705698665ec459e29d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-source","Params":"3B + 2B","Mode":"RL"}. This record does not assert independent reproduction. ### Z-Image-Turbo R-LR Score: 37.8. Result ID: 69970c705698665ec459e29a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-source","Params":"4B + 6B"}. This record does not assert independent reproduction. ### Show-o2 R-LR Score: 30.2. Result ID: 69970c705698665ec459e294. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-source","Params":"7B"}. This record does not assert independent reproduction. ### BAGEL R-LR Score: 28.6. Result ID: 69970c705698665ec459e297. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-source","Params":"14B","textual CoT reasoning":true}. This record does not assert independent reproduction. ### Janus-Pro R-LR Score: 27.2. Result ID: 69970c705698665ec459e292. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-source","Params":"7B"}. This record does not assert independent reproduction. ### OmniGen2 R-LR Score: 26.8. Result ID: 69970c705698665ec459e296. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-source","Params":"3B + 4B"}. This record does not assert independent reproduction. ### FLUX.1 \[Dev\] R-LR Score: 26.3. Result ID: 69970c705698665ec459e293. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-source","Params":"12B"}. This record does not assert independent reproduction. ### BLIP3-o R-LR Score: 18.4. Result ID: 69970c705698665ec459e295. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-source","Params":"7B + 1.4B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.