# Machine Translation (xx -\> zh) on FLORES+ (latest test) Source: [Machine Translation (xx -\> zh) on FLORES+ (latest test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:11:20.271Z Evaluation of Machine Translation performance specifically translating from various source languages (xx) to Chinese (zh) using the latest test split of the FLORES+ dataset, reporting metrics such as spBLEU and COMET. 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 | Machine Translation (xx -\> zh) | | Dataset (including any stated split) | FLORES+ latest (test) | | Metric | spBLEU | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 30 of 30 | | 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 - [spBLEU](): Higher is better; unit: PERCENT. - [COMET](): Higher is better; unit: PERCENT. ## Recorded results | Method | spBLEU | Result date | Report paper | | --- | --- | --- | --- | | Google Translate | 36.08 | Not recorded | 698eb045c97a0b3e3b5e6867 | | Gemini 3 Pro | 33.81 | Not recorded | 698eb045c97a0b3e3b5e6867 | | GemmaX2-28-9B | 33.77 | Not recorded | 698eb045c97a0b3e3b5e6867 | | Tower-Plus-9B | 33.25 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-12B | 33.16 | Not recorded | 698eb045c97a0b3e3b5e6867 | | Gemini 2.5 Pro | 33.07 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-12B | 33.06 | Not recorded | 698eb045c97a0b3e3b5e6867 | | Seed-X-Instruct-7B | 32.51 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-12B | 32.44 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-12B | 32.23 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-12B | 32.03 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-4B | 31.87 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-4B | 31.71 | Not recorded | 698eb045c97a0b3e3b5e6867 | | GPT-5 | 31.34 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-4B | 30.96 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-4B | 30.76 | Not recorded | 698eb045c97a0b3e3b5e6867 | | GemmaX2-28-2B | 30.6 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-4B | 30.53 | Not recorded | 698eb045c97a0b3e3b5e6867 | | Tower-Plus-2B | 29.53 | Not recorded | 698eb045c97a0b3e3b5e6867 | | Seed-X-PPO-7B | 29.2 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-1B | 28.31 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-1B | 26.94 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-1B | 26.76 | Not recorded | 698eb045c97a0b3e3b5e6867 | | TranslateGemma-12B | 26.75 | Not recorded | 698eb045c97a0b3e3b5e6867 | | MiLMMT-46-1B | 26.42 | Not recorded | 698eb045c97a0b3e3b5e6867 | | HY-MT1.5-7B | 23.59 | Not recorded | 698eb045c97a0b3e3b5e6867 | | TranslateGemma-4B | 23.37 | Not recorded | 698eb045c97a0b3e3b5e6867 | | Hunyuan-MT-7B | 22.98 | Not recorded | 698eb045c97a0b3e3b5e6867 | | HY-MT1.5-1.8B | 22.61 | Not recorded | 698eb045c97a0b3e3b5e6867 | | NLLB-54.5B | 20.72 | Not recorded | 698eb045c97a0b3e3b5e6867 | ### Google Translate spBLEU: 36.08. Result ID: 699637a85698665ec47a3da6. [Reporting paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### Gemini 3 Pro spBLEU: 33.81. Result ID: 699637a85698665ec47a3da7. [Reporting paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### GemmaX2-28-9B spBLEU: 33.77. Result ID: 699637a85698665ec47a3d9c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"28 languages"}. This record does not assert independent reproduction. ### Tower-Plus-9B spBLEU: 33.25. Result ID: 699637a85698665ec47a3d93. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"21 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-12B spBLEU: 33.16. Result ID: 699637a85698665ec47a3d96. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"21 languages"}. This record does not assert independent reproduction. ### Gemini 2.5 Pro spBLEU: 33.07. Result ID: 699637a85698665ec47a3da8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-12B spBLEU: 33.06. Result ID: 699637a85698665ec47a3d9a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"26 languages"}. This record does not assert independent reproduction. ### Seed-X-Instruct-7B spBLEU: 32.51. Result ID: 699637a85698665ec47a3d97. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"26 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-12B spBLEU: 32.44. Result ID: 699637a85698665ec47a3daf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-12B spBLEU: 32.23. Result ID: 699637a85698665ec47a3da5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"31 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-12B spBLEU: 32.03. Result ID: 699637a85698665ec47a3d9f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"28 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-4B spBLEU: 31.87. Result ID: 699637a85698665ec47a3d95. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"21 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-4B spBLEU: 31.71. Result ID: 699637a85698665ec47a3d99. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"26 languages"}. This record does not assert independent reproduction. ### GPT-5 spBLEU: 31.34. Result ID: 699637a85698665ec47a3da9. [Reporting paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-4B spBLEU: 30.96. Result ID: 699637a85698665ec47a3dae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-4B spBLEU: 30.76. Result ID: 699637a85698665ec47a3da4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"31 languages"}. This record does not assert independent reproduction. ### GemmaX2-28-2B spBLEU: 30.6. Result ID: 699637a85698665ec47a3d9b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"28 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-4B spBLEU: 30.53. Result ID: 699637a85698665ec47a3d9e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"28 languages"}. This record does not assert independent reproduction. ### Tower-Plus-2B spBLEU: 29.53. Result ID: 699637a85698665ec47a3d92. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"21 languages"}. This record does not assert independent reproduction. ### Seed-X-PPO-7B spBLEU: 29.2. Result ID: 699637a85698665ec47a3d98. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"26 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-1B spBLEU: 28.31. Result ID: 699637a85698665ec47a3d94. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"21 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-1B spBLEU: 26.94. Result ID: 699637a85698665ec47a3dad. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-1B spBLEU: 26.76. Result ID: 699637a85698665ec47a3da3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"31 languages"}. This record does not assert independent reproduction. ### TranslateGemma-12B spBLEU: 26.75. Result ID: 699637a85698665ec47a3dac. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### MiLMMT-46-1B spBLEU: 26.42. Result ID: 699637a85698665ec47a3d9d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"28 languages"}. This record does not assert independent reproduction. ### HY-MT1.5-7B spBLEU: 23.59. Result ID: 699637a85698665ec47a3da2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"31 languages"}. This record does not assert independent reproduction. ### TranslateGemma-4B spBLEU: 23.37. Result ID: 699637a85698665ec47a3dab. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. ### Hunyuan-MT-7B spBLEU: 22.98. Result ID: 699637a85698665ec47a3da1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"31 languages"}. This record does not assert independent reproduction. ### HY-MT1.5-1.8B spBLEU: 22.61. Result ID: 699637a85698665ec47a3da0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"31 languages"}. This record does not assert independent reproduction. ### NLLB-54.5B spBLEU: 20.72. Result ID: 699637a85698665ec47a3daa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Language Coverage":"46 languages"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.