gpt-5.4-mini remains the default recommendation for most voice agents: it has low time-to-first-token and strong instruction following at a fraction of the cost of the full models.
Quick config
Supported models
Recommendation: Start with
gpt-5.4-mini. Step up to gpt-5.6-terra or gpt-5.6-luna for newest-generation quality at moderate cost, or gpt-5.6-sol / gpt-5.5 when you need the strongest multi-step reasoning and highest output quality (financial, medical, nuanced escalation).
Key settings
Keep max_tokens short
Voice responses should be 1–3 sentences.max_tokens: 150 (about 100–120 spoken words) is appropriate for most turns. A higher cap doesn’t hurt quality but increases tail latency on long responses.
Writing prompts for voice
Prompts for voice agents differ from chat prompts:- Use imperative sentences: “Keep all responses under 3 sentences.”
- Specify spoken format: “Never use bullet points or markdown — speak in complete sentences.”
- Define handling for off-topic questions: “If asked something outside your scope, say: ‘I can only help with appointment scheduling today.’”
- Include the welcome message in the prompt or agent config, not as part of the system prompt instructions.
Function calling
GPT-4.1 and gpt-4.1-mini both support function calling. In Bolna, functions are defined in the Tools Tab and called automatically by the LLM during conversation. See Custom Function Calls for configuration.FAQ
Which GPT model should I use?
Which GPT model should I use?
Use
gpt-5.4-mini for most agents: lowest time-to-first-token and significantly lower cost per call. Step up to gpt-5.6-terra or gpt-5.6-luna for newest-generation quality at moderate cost, or gpt-5.6-sol / gpt-5.5 for the most demanding tasks (complex financial/medical, long multi-step tool chains) where quality is the top priority. Avoid gpt-5.5-pro for live voice; its latency is too high for real-time calls.Why is temperature 0.2 recommended?
Why is temperature 0.2 recommended?
Lower temperature makes the model more consistent and predictable — important for scripted outbound calls where you want the agent to follow instructions reliably. For more conversational or exploratory agents, 0.5–0.7 is fine.
How do I reduce latency?
How do I reduce latency?
Lower
max_tokens, use gpt-4.1-mini instead of larger models, and write shorter system prompts (large prompts increase prefill time). See Latency for a full breakdown.Can I use my own OpenAI API key?
Can I use my own OpenAI API key?
Yes. Connect your OpenAI account at platform.bolna.ai/auth/openai. Costs will be charged to your OpenAI account, not Bolna’s platform wallet (for the LLM component).
Related
- LLM Tab — configure LLM in the dashboard
- Anthropic Claude — alternative LLM
- Azure OpenAI — OpenAI models with enterprise data residency
- Prompting Guide — write effective prompts for voice
- Custom Function Calls — add tools to your agent