# Detailed Caption on KVBench Detailed Caption en Source: [Detailed Caption on KVBench Detailed Caption en]() Collection index: [llms.txt]() Data updated: 2026-04-27T07:09:41.059Z Evaluation of Detailed Captioning performance on the KVBench Detailed Caption (English) dataset, reporting overall and subject-specific scores across Biology, Chemistry, Geography, History, Math, and Physics. 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 | Detailed Caption | | Dataset (including any stated split) | KVBench Detailed Caption en | | Metric | Overall Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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. - [Biology Score](): Higher is better; unit: PERCENT. - [Chemistry Score](): Higher is better; unit: PERCENT. - [Geography Score](): Higher is better; unit: PERCENT. - [History Score](): Higher is better; unit: PERCENT. - [Math Score](): Higher is better; unit: PERCENT. - [Physics Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Score | Result date | Report paper | | --- | --- | --- | --- | | GPT-Image | 62.3 | Not recorded | 69eeed9955a3c32af89f97fd | | gemini-2.5-flash | 61.23 | Not recorded | 69eeed9955a3c32af89f97fd | | Seedream-4.0 | 59.67 | Not recorded | 69eeed9955a3c32af89f97fd | | gemini-3-pro-image | 57.55 | Not recorded | 69eeed9955a3c32af89f97fd | | FLUX.2-dev | 56.64 | Not recorded | 69eeed9955a3c32af89f97fd | | FLUX.2-max | 50.18 | Not recorded | 69eeed9955a3c32af89f97fd | | SD3.5-Large | 39.35 | Not recorded | 69eeed9955a3c32af89f97fd | | Qwen-Image | 37.24 | Not recorded | 69eeed9955a3c32af89f97fd | | FLUX.1-Krea-dev | 36.65 | Not recorded | 69eeed9955a3c32af89f97fd | | HiDream | 35.84 | Not recorded | 69eeed9955a3c32af89f97fd | | Z-Image | 28.9 | Not recorded | 69eeed9955a3c32af89f97fd | | FLUX.1-dev | 24.32 | Not recorded | 69eeed9955a3c32af89f97fd | | BAGEL-7B-MoT | 20.31 | Not recorded | 69eeed9955a3c32af89f97fd | | Show-o2 | 15.14 | Not recorded | 69eeed9955a3c32af89f97fd | | OmniGen2 | 14.27 | Not recorded | 69eeed9955a3c32af89f97fd | | Janus-Pro | 9.96 | Not recorded | 69eeed9955a3c32af89f97fd | ### GPT-Image Overall Score: 62.3. Result ID: 69ef0a9a47f34ce924d4bd23. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Close-source Models"}. This record does not assert independent reproduction. ### gemini-2.5-flash Overall Score: 61.23. Result ID: 69ef0a9a47f34ce924d4bd21. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Source Category":"Close-source Models"}. This record does not assert independent reproduction. ### Seedream-4.0 Overall Score: 59.67. Result ID: 69ef0a9a47f34ce924d4bd22. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Source Category":"Close-source Models"}. This record does not assert independent reproduction. ### gemini-3-pro-image Overall Score: 57.55. Result ID: 69ef0a9a47f34ce924d4bd24. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Close-source Models"}. This record does not assert independent reproduction. ### FLUX.2-dev Overall Score: 56.64. Result ID: 69ef0a9a47f34ce924d4bd2c. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### FLUX.2-max Overall Score: 50.18. Result ID: 69ef0a9a47f34ce924d4bd20. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Close-source Models"}. This record does not assert independent reproduction. ### SD3.5-Large Overall Score: 39.35. Result ID: 69ef0a9a47f34ce924d4bd27. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### Qwen-Image Overall Score: 37.24. Result ID: 69ef0a9a47f34ce924d4bd2e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### FLUX.1-Krea-dev Overall Score: 36.65. Result ID: 69ef0a9a47f34ce924d4bd29. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### HiDream Overall Score: 35.84. Result ID: 69ef0a9a47f34ce924d4bd1f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Source Category":"Close-source Models"}. This record does not assert independent reproduction. ### Z-Image Overall Score: 28.9. Result ID: 69ef0a9a47f34ce924d4bd2d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### FLUX.1-dev Overall Score: 24.32. Result ID: 69ef0a9a47f34ce924d4bd26. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### BAGEL-7B-MoT Overall Score: 20.31. Result ID: 69ef0a9a47f34ce924d4bd2b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### Show-o2 Overall Score: 15.14. Result ID: 69ef0a9a47f34ce924d4bd25. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### OmniGen2 Overall Score: 14.27. Result ID: 69ef0a9a47f34ce924d4bd2a. [Reporting paper]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. ### Janus-Pro Overall Score: 9.96. Result ID: 69ef0a9a47f34ce924d4bd28. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Source Category":"Open-source Models"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.