{"benchmark_id":"covost2","name":"CoVoST2","parent_benchmark":null,"categories":["speech_to_text","language","audio"],"modality":"audio","multilingual":true,"max_score":1.0,"language":"en","description":"CoVoST 2 is a large-scale multilingual speech translation corpus derived from Common Voice, covering translations from 21 languages into English and from English into 15 languages. The dataset contains 2,880 hours of speech with 78K speakers for speech translation research.","paper_link":"https://arxiv.org/abs/2007.10310","implementation_link":null,"verified":false,"created_at":"2026-05-07T16:53:22.554410+00:00","updated_at":"2026-07-05T18:27:46.009083+00:00","statistics":{"total_models":4,"average_score":0.392,"min_score":0.384,"max_score":0.407,"score_stddev":0.010614455552060413,"verified_count":0,"self_reported_count":4},"child_benchmarks":[],"linked_dataset":null,"models":[{"rank":1,"model_id":"nova-2-omni","model_name":"Nova 2 Omni","organization_id":"amazon","organization_name":"Amazon","organization_country":"US","score":0.407,"normalized_score":0.407,"verified":false,"self_reported":true,"self_reported_source":"https://cdn.amazon.science/c5/3d/84514a224666b5be6de4b43ef4aa/nova-2-0-technical-report2.pdf","analysis_method":"BLEU (X-to-English)","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-12-02","param_count":null,"is_open_source":false,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":2,"model_id":"gemini-2.0-flash","model_name":"Gemini 2.0 Flash","organization_id":"google","organization_name":"Google","organization_country":"US","score":0.392,"normalized_score":0.392,"verified":false,"self_reported":true,"self_reported_source":"https://blog.google/technology/google-deepmind/google-gemini-ai-update-december-2024/","analysis_method":"Automatic speech translation (BLEU score) across 21 languages","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2024-12-01","param_count":null,"is_open_source":false,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":3,"model_id":"gemma-4-12b-it","model_name":"Gemma 4 12B","organization_id":"google","organization_name":"Google","organization_country":"US","score":0.385,"normalized_score":0.385,"verified":false,"self_reported":true,"self_reported_source":"https://huggingface.co/google/gemma-4-12B-it","analysis_method":"BLEU","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2026-05-23","param_count":11959730224,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":4,"model_id":"gemini-2.0-flash-lite","model_name":"Gemini 2.0 Flash-Lite","organization_id":"google","organization_name":"Google","organization_country":"US","score":0.384,"normalized_score":0.384,"verified":false,"self_reported":true,"self_reported_source":"https://ai.google.dev/gemini-api/docs/models","analysis_method":"Automatic speech translation (BLEU score) across 21 languages","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-02-05","param_count":null,"is_open_source":false,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null}]}