{
  "model_name": "qwen3.5-omni-plus-realtime",
  "model_organization": "Qwen",
  "submitting_organization": "Sierra",
  "submission_date": "2026-08-06",
  "submission_type": "standard",
  "modality": "voice",
  "contact_info": {
    "email": "keshav@sierra.ai",
    "name": "Sierra Research Team"
  },
  "results": {
    "retail": {
      "pass_1": 46.49122807017544,
      "cost": 0.580765906741573
    },
    "airline": {
      "pass_1": 54.0,
      "cost": 0.5397773999999999
    },
    "telecom": {
      "pass_1": 60.526315789473685,
      "cost": 1.467414694736842
    }
  },
  "is_new": true,
  "trajectories_available": true,
  "trajectory_files": {
    "airline": "airline_regular_qwen",
    "retail": "retail_regular_qwen",
    "telecom": "telecom_regular_qwen"
  },
  "methodology": {
    "evaluation_date": "2026-08-05",
    "tau2_bench_version": "1.0.1",
    "user_simulator": "v1.0",
    "notes": "Full-duplex audio-native evaluation using regular speech complexity (realistic audio conditions). 1 trial per task.",
    "verification": {
      "modified_prompts": false,
      "omitted_questions": false
    }
  },
  "voice_config": {
    "provider": "qwen",
    "model": "qwen3.5-omni-plus-realtime",
    "tick_duration_seconds": 0.2,
    "max_steps_seconds": 1200.0,
    "user_tts_provider": "elevenlabs/eleven_v3"
  },
  "interaction_metrics": {
    "version": "1.0",
    "config": {
      "tick_duration_sec": 0.2,
      "no_yield_window_sec": 2.0,
      "backchannel_yield_window_sec": 1.0,
      "vocal_tic_yield_window_sec": 1.0,
      "non_directed_yield_window_sec": 1.0,
      "vocal_tic_response_window_sec": 2.0,
      "non_directed_response_window_sec": 2.0
    },
    "domains": {
      "airline": {
        "response_latency_mean": 1.7473451327433622,
        "yield_latency_mean": 0.9282352941176472,
        "response_rate": 0.9847494553376906,
        "yield_rate": 0.8374384236453202,
        "agent_interruption_rate": 0.41830065359477125,
        "selectivity_backchannel": 0.926829268292683,
        "selectivity_vocal_tic": 0.3770491803278688,
        "selectivity_non_directed": 0.18367346938775508,
        "counts": {
          "n_simulations": 50,
          "response_total": 459,
          "yield_total": 406,
          "backchannel_total": 41,
          "vocal_tic_total": 61,
          "non_directed_total": 49,
          "agent_interrupts_count": 192
        }
      },
      "retail": {
        "response_latency_mean": 1.706824925816023,
        "yield_latency_mean": 0.9324855491329481,
        "response_rate": 0.9931237721021611,
        "yield_rate": 0.9124472573839663,
        "agent_interruption_rate": 0.35952848722986247,
        "selectivity_backchannel": 0.7439024390243902,
        "selectivity_vocal_tic": 0.37755102040816324,
        "selectivity_non_directed": 0.15596330275229353,
        "counts": {
          "n_simulations": 114,
          "response_total": 1018,
          "yield_total": 948,
          "backchannel_total": 82,
          "vocal_tic_total": 98,
          "non_directed_total": 109,
          "agent_interrupts_count": 366
        }
      },
      "telecom": {
        "response_latency_mean": 1.6407796101948993,
        "yield_latency_mean": 0.9368817710786622,
        "response_rate": 0.9977561705310396,
        "yield_rate": 0.9558757316524088,
        "agent_interruption_rate": 0.30553477935676887,
        "selectivity_backchannel": 0.8493150684931507,
        "selectivity_vocal_tic": 0.33624454148471616,
        "selectivity_non_directed": 0.09569377990430628,
        "counts": {
          "n_simulations": 114,
          "response_total": 2674,
          "yield_total": 2221,
          "backchannel_total": 73,
          "vocal_tic_total": 229,
          "non_directed_total": 209,
          "agent_interrupts_count": 817
        }
      }
    },
    "overall": {
      "response_latency_mean": 1.6983165562514282,
      "yield_latency_mean": 0.9325342047764192,
      "response_rate": 0.9918764659902971,
      "yield_rate": 0.9019204708938985,
      "agent_interruption_rate": 0.36112130672713416,
      "selectivity_backchannel": 0.8400155919367412,
      "selectivity_vocal_tic": 0.3636149140735827,
      "selectivity_non_directed": 0.14511018401478495,
      "counts": {
        "agent_interrupts_count": 1375,
        "backchannel_total": 196,
        "n_simulations": 278,
        "non_directed_total": 367,
        "response_total": 4151,
        "vocal_tic_total": 388,
        "yield_total": 3575
      }
    }
  },
  "model_release": {
    "release_date": "2026-03-30",
    "announcement_url": "https://www.alibabacloud.com/help/en/model-studio/qwen-omni",
    "announcement_title": "Qwen-Omni Models | Alibaba Cloud Model Studio"
  }
}