GPT-Live 1 + GPT-6 Sol vs GPT Realtime 2.1

They are level on reliability. GPT-Live 1 + GPT-6 Sol handled 23 of 70 real callers well on all three runs, and GPT Realtime 2.1 23. A caller only counts when every run of the call scored 80 or more.

Reliability does not separate them, so pick on the metric your calls need most. GPT-Live 1 + GPT-6 Sol is ahead on ELO rating, speaking score, task handling score, calls handled well, calls fully resolved and cost, and GPT Realtime 2.1 on hearing score, turn taking score, calls where the agent went quiet and median latency. The two tie on reliability.

GPT-Live 1 + GPT-6 Sol, by 1.9 points higher. GPT-Live 1 + GPT-6 Sol fully resolved 36.2% of calls and GPT Realtime 2.1 34.3%. A call counts only when every step the caller needed was done.

No. GPT Realtime 2.1 starts replying 0.11s sooner at the median. GPT-Live 1 + GPT-6 Sol takes 2.27s at the median and 2.72s at p90; GPT Realtime 2.1 takes 2.16s and 2.70s. Both are measured on the call audio, from the caller finishing to the agent starting to reply.

GPT Realtime 2.1, by 1.4 points. The agent went quiet for 10 seconds or more on 13.3% of GPT-Live 1 + GPT-6 Sol calls and 11.9% of GPT Realtime 2.1 calls.

GPT-Live 1 + GPT-6 Sol is $0.030 cheaper a minute. GPT-Live 1 + GPT-6 Sol costs $0.055 a minute of call and GPT Realtime 2.1 $0.085, at each vendor's published rates.

Cekura ran both as the whole agent with the same prompt, tools and phone connection. Both took the same 70 recorded real human callers on live calls, three times each. Each call is graded on 25 skills across Hearing, Turn Taking, Task Handling and Speaking, and passes at a score of 80 out of 100.