Skip to content

Scenario guide

Best AI models for Realtime Chat / Voice

As of October 25, 2026, the highest-ranked model for Realtime Chat / Voice is GPT-5.6 Terra Pro, with a scenario score of 83.8 out of 100. The score blends this scenario's benchmark weights with its cost weight.

A streaming consumer chat or voice assistant. Speed and time-to-first-token matter as much as raw quality — a slightly less smart model that responds instantly often beats a frontier model that pauses to think.

Rankings use the same scenario weights and cost blending as the interactive leaderboard on AI Model Analyzer. Data is min-max normalised per benchmark; missing scores are skipped without penalty.

  1. 1
    GPT-5.6 Terra Pro
    OpenAI
    Score 83.8Q 96.7In $1.00/M
  2. 2
    Claude Fable 5.1
    Anthropic
    Score 76.2Q 96.4In $10.00/M
  3. 3
    Claude Fable 5
    Anthropic
    Score 72.0Q 90.4In $10.00/M
  4. 4
    Claude Haiku 5.5
    Anthropic
    Score 70.2Q 65.6In $0.10/M
  5. 5
    GLM 5
    Zhipu AI (GLM)
    Score 68.8Q 73.7In $1.00/M
  6. 6
    Llama 3.3 70B Instruct
    Meta
    Score 64.2Q 56.0In $0.10/M
  7. 7
    GPT-6.1 Sol Pro
    OpenAI
    Score 63.1Q 66.7In $1.00/M
  8. 8
    DeepSeek V4 Flash 0423
    DeepSeek
    Score 62.9Q 52.8In $0.09/M
  9. 9
    DeepSeek V4.1 Flash
    DeepSeek
    Score 62.4Q 55.4In $0.02/M
  10. 10
    GLM 5.3 Flash
    Zhipu AI (GLM)
    Score 62.2Q 55.2In $0.15/M
  11. 11
    MiMo-V2.6-Flash
    Xiaomi (MiMo)
    Score 62.2Q 53.9In $0.14/M
  12. 12
    MiMo-V2.5
    Xiaomi (MiMo)
    Score 61.7Q 52.5In $0.12/M
  13. 13
    Qwen3.5-Flash
    Alibaba (Qwen)
    Score 61.6Q 50.8In $0.07/M
  14. 14
    DeepSeek V4 Pro 0423
    DeepSeek
    Score 61.1Q 54.3In $0.21/M
  15. 15
    Gemma 4 31B
    Google
    Score 61.0Q 54.0In $0.20/M
Open interactive leaderboard Build custom weightsHome