### Answer For Retell AI voice agents, the available LLM options include gpt-4.1, gpt-4.1-mini, and gpt-4.1-nano among others [(1) ]() [(2) ]() . The documentation shows that if not set, the default is gpt-4.1 [(2) ]() . When selecting an LLM model, you can configure it using the

model

parameter [(1) ]() [(2) ]() :

javascript
model: "gpt-4.1"

The

model_temperature

parameter controls randomness, with lower values (like 0) being more deterministic—which is recommended for tool calling :

javascript
model_temperature: 0

For voice-specific optimization, Retell AI also offers speech-to-speech models like gpt-realtime-1.5 via the

s2s_model

parameter . Note that you can only set either

model

or

s2s_model

, not both . The best choice depends on your use case. For complex prompt following and tool calling, gpt-4.1 with low temperature is recommended . For conversational applications prioritizing natural speech patterns, consider the gpt-realtime-1.5 speech-to-speech model .

​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[2] Update Retell LLM](https://docs.retellai.com/api-references/update-retell-llm) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer \Based on the available models in Retell AI, here's what the platform currently supports for voice agents [(1) ]() [(2) ]() : Text LLM Models: - gpt-4.1, gpt-4.1-mini, gpt-4.1-nano - gpt-5, gpt-5-mini, gpt-5-nano - gpt-5.1, gpt-5.2, gpt-5.4, gpt-5.4-mini, gpt-5.4-nano - claude-4.5-sonnet, claude-4.6-sonnet, claude-4.5-haiku - gemini-2.5-flash, gemini-2.5-flash-lite, gemini-3.0-flash Speech-to-Speech Models: - gpt-realtime-1.5, gpt-realtime, gpt-realtime-mini Recent Release: GPT-5.1 was recently released and provides more conversational, humanlike responses, answers up to 5× faster on most questions, follows instructions more accurately, and handles complex, function-heavy workflows more reliably [(3) ]() . For voice agents specifically, the documentation indicates that if

model

is not set, it defaults to gpt-4.1 [(1) ]() [(2) ]() . For tool calling, a lower

model_temperature

value is recommended . The "best" model depends on your specific needs—intelligence, latency, or cost. Based on the changelog, GPT-5.1 appears to be optimized for natural conversations with improved speed and reliability [(3) ]() .

​ To learn more, see the sources I considered: [[1] Update Retell LLM](https://docs.retellai.com/api-references/update-retell-llm) [[2] Create Retell LLM](https://docs.retellai.com/api-references/create-retell-llm) [[3] 🚀 New Model Release: GPT-5.1 for More Natural Conversations](https://community.retellai.com/t/new-model-release-gpt-5-1-for-more-natural-conversations/164) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer Based on the available sources, here's the pricing information for different LLM models on Retell AI [(1) ]() [(2) ]() : LLM Agent Pricing (per minute) : [(1) ]() - Retell LLM - GPT 3.5: $0.02/min - Retell LLM - GPT 4.0: $0.2/min - Custom LLM: No charge GPT-4.1 Family Pricing (per minute) : [(2) ]() - GPT-4.1 nano: $0.004/min - GPT-4o mini: $0.006/min - GPT-4.1 mini: $0.016/min - GPT-4.1: $0.045/min - GPT-4o: $0.050/min Additional Cost Considerations: Starting June 1st, Retell AI implemented token-based LLM pricing . Prompts up to 3,500 tokens follow the flat rate above . Prompts exceeding 3,500 tokens scale proportionally . For example with GPT-4.1: if your prompt has 4,500 tokens, the cost would be: 0.07 + (4500 / 3500 × 0.045) = 0.1279 per minute . Fast Tier (Premium Performance) : [(3) ]() If you enable Fast Tier for better latency and consistency, pricing is 1.5x the standard rate for your selected model . For the most cost-effective option, GPT-4.1 nano at $0.004/min offers the lowest price [(2) ]() , while GPT-4.1 at $0.045/min provides higher intelligence .

​ To learn more, see the sources I considered: [[1] New LLM Model Options, Modular Pricing, And More](https://www.retellai.com/changelog/new-llm-model-options-modular-pricing-and-more) [[2] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[3] Configure LLM Options](https://docs.retellai.com/build/llm-options) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer Based on the pricing and performance information, here's what the sources show: For Cost-Sensitive Use Cases: - gpt-4.1-nano at $0.004/min offers the cheapest option [(1) ]() - gpt-5-nano is also designed for cost-sensitive applications For Balanced Performance: - gpt-4.1-mini at $0.016/min provides a strong balance between cost and intelligence - gpt-5-mini offers lower cost with strong intelligence Temperature Configuration: For tool calling and function execution, set a low temperature value [(2) ]() [(3) ]() [(4) ]() :

javascript
model_temperature: 0

[(2) ]() , [(3) ]() , [(4) ]() Key Considerations from the Documentation: The sources indicate that for tool calling, a lower

model_temperature

value is recommended [(2) ]() [(3) ]() [(4) ]() . Temperature controls randomness - values between 0.0-0.3 are highly consistent and deterministic, which is best for function calling, data collection, and technical support [(2) ]() . For appointment booking specifically, the documentation recommends using 0.1-0.3 for accurate data capture . Cost Optimization: Prompts up to 3,500 tokens follow the flat rate pricing [(1) ]() . If your prompts exceed this, costs scale proportionally . The default model when not specified is gpt-4.1 [(3) ]() [(4) ]() . **

​ To learn more, see the sources I considered: [[1] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[2] Configure LLM Options](https://docs.retellai.com/build/llm-options) [[3] Get Retell LLM](https://docs.retellai.com/api-references/get-retell-llm) [[4] Update Retell LLM](https://docs.retellai.com/api-references/update-retell-llm) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer gpt-4.1 is absolutely a strong choice - it's actually the default model when not specified [(1) ]() [(2) ]() [(3) ]() . At $0.045/min [(4) ]() , it offers the highest intelligence in the GPT-4.1 family . Why gpt-4.1 is excellent: - Highest intelligence among GPT-4.1 models - Default model for Retell AI [(1) ]() [(2) ]() [(3) ]() - Enhanced reasoning capabilities [(5) ]() - Extended memory and context (up to 1 million tokens in top-tier) - Natural conversation flow with reduced latency The trade-off is cost: - gpt-4.1: $0.045/min [(4) ]() - gpt-4.1-mini: $0.016/min (about 3x cheaper) - gpt-4.1-nano: $0.004/min (about 11x cheaper) If your use case requires maximum intelligence, complex reasoning, or handling sophisticated multi-step problems, gpt-4.1 is worth the cost [(5) ]() . For simpler tasks like basic appointment booking or straightforward data collection, the mini or nano versions offer significant savings while still performing well [(4) ]() . The intelligence ranking is: GPT 4.1 > GPT 4o mini > GPT 4o > GPT 4.1 mini > GPT 4.1 nano .

​ To learn more, see the sources I considered: [[1] Get Retell LLM](https://docs.retellai.com/api-references/get-retell-llm) [[2] Create Retell LLM](https://docs.retellai.com/api-references/create-retell-llm) [[3] Update Retell LLM](https://docs.retellai.com/api-references/update-retell-llm) [[4] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[5] Smarter AI Phone Calls with ChatGPT-4.1 | Retell AI | Retell AI](https://www.retellai.com/blog/chatgpt-4-1-ai-phone-calling-retell) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.