Pipecat PhoneLLM vs Ultravox Realtime API

Pipecat PhoneLLM

PhoneLLM Mini is callable per token on Workers AI at $0.06 input and $0.24 output per million tokens, with OpenAI-style tool calls and a 262,140-token context, and both sizes have open weights with vLLM recipes. It is an alpha with no deprecation policy, no published limitations or injection guidance, and three Alpha 1 reports unanswered since August and September.

Ultravox Realtime API

Flat $0.05 a minute including model and built-in voices. No MCP server.

Pipecat PhoneLLM

1. Send temperature: 0 on every call. The repositories' generation_config.json samples by default, and Max's sets temperature 1.0
2. For Max, serve with --reasoning-parser nemotron_v3 and send chat_template_kwargs with enable_thinking: true on every request, or the reply lands in reasoning
3. For Mini, serve without a reasoning parser and keep enable_thinking false, as Cloudflare and the card say, or tool calls can be filed under reasoning
4. Confirm bookings, payments and transfers in your own handler before acting on a tool call, since caller speech reaches the model directly
5. Pin the model ID including alpha-2. Daily states no notice period, and a later alpha may replace it

Ultravox Realtime API

1. Honour Retry-After on 429 and 503 when creating calls
2. Pin the model string (ultravox-v0.7 or ultravox-v0.6) rather than relying on the default
3. Hand clients the joinUrl, never the API key
4. Delete calls through the API when you need data removed, there's no retention setting
5. Use call stages for multi-step flows instead of one long prompt