{"benchmark_id":"tau2-airline","name":"Tau2 Airline","parent_benchmark":null,"categories":["reasoning","communication","tool_calling"],"modality":"text","multilingual":false,"max_score":1.0,"language":"en","description":"TAU2 airline domain benchmark for evaluating conversational agents in dual-control environments where both AI agents and users interact with tools in airline customer service scenarios. Tests agent coordination, communication, and ability to guide user actions in tasks like flight booking, modifications, cancellations, and refunds.","paper_link":"https://arxiv.org/abs/2506.07982","implementation_link":null,"verified":false,"created_at":"2026-05-07T16:53:26.430077+00:00","updated_at":"2026-07-05T18:28:02.695080+00:00","statistics":{"total_models":23,"average_score":0.5976304347826086,"min_score":0.44,"max_score":0.765,"score_stddev":0.08433701567780742,"verified_count":0,"self_reported_count":23},"child_benchmarks":[],"linked_dataset":null,"models":[{"rank":1,"model_id":"longcat-flash-thinking-2601","model_name":"LongCat-Flash-Thinking-2601","organization_id":"meituan","organization_name":"Meituan","organization_country":"CN","score":0.765,"normalized_score":0.765,"verified":false,"self_reported":true,"self_reported_source":"https://huggingface.co/meituan-longcat/LongCat-Flash-Thinking-2601","analysis_method":"Avg@4","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2026-01-14","param_count":560000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":2,"model_id":"nova-2-omni","model_name":"Nova 2 Omni","organization_id":"amazon","organization_name":"Amazon","organization_country":"US","score":0.688,"normalized_score":0.688,"verified":false,"self_reported":true,"self_reported_source":"https://cdn.amazon.science/c5/3d/84514a224666b5be6de4b43ef4aa/nova-2-0-technical-report2.pdf","analysis_method":"Verified avg@3","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":3,"model_id":"longcat-flash-thinking","model_name":"LongCat-Flash-Thinking","organization_id":"meituan","organization_name":"Meituan","organization_country":"CN","score":0.675,"normalized_score":0.675,"verified":false,"self_reported":true,"self_reported_source":"https://github.com/meituan-longcat/LongCat-Flash-Thinking","analysis_method":"Mean@4","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-09-22","param_count":560000000000,"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":"gpt-5.1-thinking-2025-11-12","model_name":"GPT-5.1 Thinking","organization_id":"openai","organization_name":"OpenAI","organization_country":"US","score":0.67,"normalized_score":0.67,"verified":false,"self_reported":true,"self_reported_source":"https://openai.com/index/gpt-5-1/","analysis_method":"GPT-5.1 Thinking with adaptive reasoning - Function calling benchmark (airline domain).","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-11-12","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":4,"model_id":"gpt-5.1-2025-11-13","model_name":"GPT-5.1","organization_id":"openai","organization_name":"OpenAI","organization_country":"US","score":0.67,"normalized_score":0.67,"verified":false,"self_reported":true,"self_reported_source":"https://openai.com/index/gpt-5-1/","analysis_method":"GPT-5.1 with adaptive reasoning - Function calling benchmark (airline domain).","verification_date":null,"provider_id":"openai","input_cost_per_million":1.25,"output_cost_per_million":10.0,"context_window":400000,"announcement_date":"2025-11-13","param_count":null,"is_open_source":false,"is_new":false,"best_latency":2.0,"latency_provider":"OpenAI","best_throughput":100.0,"throughput_provider":"OpenAI","context_provider":"OpenAI"},{"rank":4,"model_id":"gpt-5.1-instant-2025-11-12","model_name":"GPT-5.1 Instant","organization_id":"openai","organization_name":"OpenAI","organization_country":"US","score":0.67,"normalized_score":0.67,"verified":false,"self_reported":true,"self_reported_source":"https://openai.com/index/gpt-5-1/","analysis_method":"GPT-5.1 with adaptive reasoning - Function calling benchmark (airline domain).","verification_date":null,"provider_id":"openai","input_cost_per_million":1.25,"output_cost_per_million":10.0,"context_window":400000,"announcement_date":"2025-11-12","param_count":null,"is_open_source":false,"is_new":false,"best_latency":2.0,"latency_provider":"OpenAI","best_throughput":100.0,"throughput_provider":"OpenAI","context_provider":"OpenAI"},{"rank":7,"model_id":"nova-2-pro","model_name":"Nova 2 Pro","organization_id":"amazon","organization_name":"Amazon","organization_country":"US","score":0.652,"normalized_score":0.652,"verified":false,"self_reported":true,"self_reported_source":"https://cdn.amazon.science/c5/3d/84514a224666b5be6de4b43ef4aa/nova-2-0-technical-report2.pdf","analysis_method":"Verified avg@3","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":8,"model_id":"nova-2-lite","model_name":"Nova 2 Lite","organization_id":"amazon","organization_name":"Amazon","organization_country":"US","score":0.648,"normalized_score":0.648,"verified":false,"self_reported":true,"self_reported_source":"https://cdn.amazon.science/c5/3d/84514a224666b5be6de4b43ef4aa/nova-2-0-technical-report2.pdf","analysis_method":"Verified avg@3","verification_date":null,"provider_id":"bedrock","input_cost_per_million":0.3,"output_cost_per_million":2.5,"context_window":1000000,"announcement_date":"2025-12-02","param_count":null,"is_open_source":false,"is_new":false,"best_latency":null,"latency_provider":"Bedrock","best_throughput":null,"throughput_provider":"Bedrock","context_provider":"Bedrock"},{"rank":8,"model_id":"o3-2025-04-16","model_name":"o3","organization_id":"openai","organization_name":"OpenAI","organization_country":"US","score":0.648,"normalized_score":0.648,"verified":false,"self_reported":true,"self_reported_source":"https://openai.com/index/gpt-5/","analysis_method":"OpenAI o3 with thinking mode - Function calling benchmark (airline domain).","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-04-16","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":10,"model_id":"claude-haiku-4-5-20251001","model_name":"Claude Haiku 4.5","organization_id":"anthropic","organization_name":"Anthropic","organization_country":"US","score":0.636,"normalized_score":0.636,"verified":false,"self_reported":true,"self_reported_source":"https://www.anthropic.com/news/claude-haiku-4-5","analysis_method":"standard","verification_date":null,"provider_id":"anthropic","input_cost_per_million":1.0,"output_cost_per_million":5.0,"context_window":200000,"announcement_date":"2025-10-15","param_count":null,"is_open_source":false,"is_new":false,"best_latency":0.3,"latency_provider":"Anthropic","best_throughput":100.0,"throughput_provider":"Anthropic","context_provider":"Anthropic"},{"rank":11,"model_id":"gpt-5-2025-08-07","model_name":"GPT-5","organization_id":"openai","organization_name":"OpenAI","organization_country":"US","score":0.626,"normalized_score":0.626,"verified":false,"self_reported":true,"self_reported_source":"https://openai.com/index/introducing-gpt-5-for-developers/","analysis_method":"GPT-5 - Function calling benchmark (airline domain).","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-08-07","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":12,"model_id":"qwen3-next-80b-a3b-thinking","model_name":"Qwen3-Next-80B-A3B-Thinking","organization_id":"qwen","organization_name":"Alibaba Cloud / Qwen Team","organization_country":"CN","score":0.605,"normalized_score":0.605,"verified":false,"self_reported":true,"self_reported_source":"https://qwenlm.github.io/blog/qwen3-next/","analysis_method":null,"verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-09-10","param_count":80000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":13,"model_id":"longcat-flash-chat","model_name":"LongCat-Flash-Chat","organization_id":"meituan","organization_name":"Meituan","organization_country":"CN","score":0.58,"normalized_score":0.58,"verified":false,"self_reported":true,"self_reported_source":"https://huggingface.co/meituan-longcat/LongCat-Flash-Chat","analysis_method":"avg@4","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-08-29","param_count":560000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":13,"model_id":"qwen3-235b-a22b-thinking-2507","model_name":"Qwen3-235B-A22B-Thinking-2507","organization_id":"qwen","organization_name":"Alibaba Cloud / Qwen Team","organization_country":"CN","score":0.58,"normalized_score":0.58,"verified":false,"self_reported":true,"self_reported_source":"https://qwenlm.github.io/blog/qwen3-thinking/","analysis_method":null,"verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-07-25","param_count":235000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":13,"model_id":"longcat-flash-lite","model_name":"LongCat-Flash-Lite","organization_id":"meituan","organization_name":"Meituan","organization_country":"CN","score":0.58,"normalized_score":0.58,"verified":false,"self_reported":true,"self_reported_source":"https://huggingface.co/meituan-longcat/LongCat-Flash-Lite","analysis_method":"avg@8","verification_date":null,"provider_id":"meituan","input_cost_per_million":0.1,"output_cost_per_million":0.4,"context_window":256000,"announcement_date":"2026-02-05","param_count":68500000000,"is_open_source":true,"is_new":false,"best_latency":1.5,"latency_provider":"Meituan","best_throughput":500.0,"throughput_provider":"Meituan","context_provider":"Meituan"},{"rank":16,"model_id":"kimi-k2-instruct","model_name":"Kimi K2 Instruct","organization_id":"moonshotai","organization_name":"Moonshot AI","organization_country":"CN","score":0.565,"normalized_score":0.565,"verified":false,"self_reported":true,"self_reported_source":"https://moonshotai.github.io/Kimi-K2/","analysis_method":"Avg@4","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-07-11","param_count":1000000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":16,"model_id":"kimi-k2-instruct-0905","model_name":"Kimi K2-Instruct-0905","organization_id":"moonshotai","organization_name":"Moonshot AI","organization_country":"CN","score":0.565,"normalized_score":0.565,"verified":false,"self_reported":true,"self_reported_source":"https://moonshotai.github.io/Kimi-K2/","analysis_method":"Avg@4","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-09-05","param_count":1000000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":18,"model_id":"nemotron-3-super-120b-a12b","model_name":"Nemotron 3 Super (120B A12B)","organization_id":"nvidia","organization_name":"NVIDIA","organization_country":"US","score":0.5625,"normalized_score":0.5625,"verified":false,"self_reported":true,"self_reported_source":"https://build.nvidia.com/nvidia/nemotron-3-super-120b-a12b/modelcard","analysis_method":null,"verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2026-03-11","param_count":120000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":19,"model_id":"mercury-2","model_name":"Mercury 2","organization_id":"inceptionlabs","organization_name":"Inception","organization_country":"US","score":0.53,"normalized_score":0.53,"verified":false,"self_reported":true,"self_reported_source":"https://www.inceptionlabs.ai/blog/introducing-mercury-2","analysis_method":"Accuracy","verification_date":null,"provider_id":"inception","input_cost_per_million":0.25,"output_cost_per_million":0.75,"context_window":128000,"announcement_date":"2026-02-24","param_count":null,"is_open_source":false,"is_new":false,"best_latency":1.7,"latency_provider":"Inception","best_throughput":1009.0,"throughput_provider":"Inception","context_provider":"Inception"},{"rank":20,"model_id":"nemotron-3-nano-30b-a3b","model_name":"Nemotron 3 Nano (30B A3B)","organization_id":"nvidia","organization_name":"NVIDIA","organization_country":"US","score":0.48,"normalized_score":0.48,"verified":false,"self_reported":true,"self_reported_source":"https://build.nvidia.com/nvidia/nemotron-3-nano-30b-a3b/modelcard","analysis_method":null,"verification_date":null,"provider_id":"deepinfra","input_cost_per_million":0.06,"output_cost_per_million":0.24,"context_window":262144,"announcement_date":"2025-12-15","param_count":32000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":"DeepInfra","best_throughput":null,"throughput_provider":"DeepInfra","context_provider":"DeepInfra"},{"rank":21,"model_id":"qwen3-next-80b-a3b-instruct","model_name":"Qwen3-Next-80B-A3B-Instruct","organization_id":"qwen","organization_name":"Alibaba Cloud / Qwen Team","organization_country":"CN","score":0.455,"normalized_score":0.455,"verified":false,"self_reported":true,"self_reported_source":"https://qwenlm.github.io/blog/qwen3-next/","analysis_method":null,"verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-09-10","param_count":80000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null},{"rank":21,"model_id":"gpt-4o-2024-08-06","model_name":"GPT-4o","organization_id":"openai","organization_name":"OpenAI","organization_country":"US","score":0.455,"normalized_score":0.455,"verified":false,"self_reported":true,"self_reported_source":"https://openai.com/index/gpt-5/","analysis_method":"GPT-4o without thinking mode - Function calling benchmark (airline domain).","verification_date":null,"provider_id":"openai","input_cost_per_million":2.5,"output_cost_per_million":10.0,"context_window":128000,"announcement_date":"2024-08-06","param_count":null,"is_open_source":false,"is_new":false,"best_latency":0.5,"latency_provider":"OpenAI","best_throughput":132.0,"throughput_provider":"OpenAI","context_provider":"OpenAI"},{"rank":23,"model_id":"qwen3-235b-a22b-instruct-2507","model_name":"Qwen3-235B-A22B-Instruct-2507","organization_id":"qwen","organization_name":"Alibaba Cloud / Qwen Team","organization_country":"CN","score":0.44,"normalized_score":0.44,"verified":false,"self_reported":true,"self_reported_source":"https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507","analysis_method":"Accuracy","verification_date":null,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"context_window":null,"announcement_date":"2025-07-22","param_count":235000000000,"is_open_source":true,"is_new":false,"best_latency":null,"latency_provider":null,"best_throughput":null,"throughput_provider":null,"context_provider":null}]}