☎
AI voice SDR
SalesAn outbound/inbound voice agent that books meetings. A three-model pipeline: speech-to-text → reasoning → text-to-speech, all on low-latency infra.
The routing recipe
Job typeGeneral chat / assistantPolicyFastest that passesConstraintsgroq
routes to
GPT-OSS 20B groq
groq:openai/gpt-oss-20b
87.9
match score
77.7
task fit
75.4
cost
96.4
speed
Why this model from benchmarks & capabilities
- "General chat / assistant" leans hardest on instruction; GPT-OSS 20B scores 78/100 there — #18 of 32.
- Strength on this class of work shows up on LMArena Elo, MT-Bench, IFEval, where OpenAI (open-weight) models are competitive.
- Cost: ~$0.112/Mtok blended ($0.03 in / $0.14 out). Speed: ~1000 tok/s on Groq LPU.
- Chosen over GPT-OSS 120B for cheaper & faster.
| Capability the job needs | Weight | GPT-OSS 20B | Catalog rank |
|---|---|---|---|
| instruction | 0.80 | #18 of 32 | |
| knowledge | 0.60 | #21 of 32 | |
| speed | 0.60 | #1 of 32 |
Relevant benchmarksLMArena Elo MT-Bench IFEval
EconomicsCost: ~$0.112/Mtok blended ($0.03 in / $0.14 out).
Speed: ~1000 tok/s on Groq LPU. vs runner-upChosen over GPT-OSS 120B for cheaper & faster. (GPT-OSS 120B, score 85.1)
Speed: ~1000 tok/s on Groq LPU. vs runner-upChosen over GPT-OSS 120B for cheaper & faster. (GPT-OSS 120B, score 85.1)
Model pipeline
Listen (STT)
groq:whisper-large-v3-turbo
$0.04/hr audio, fastest transcription
Think (LLM)
groq:openai/gpt-oss-20b
fastest Groq chat model above the quality floor
Speak (TTS)
groq:canopylabs/orpheus-3b-0.1-ft-en
low-latency English voice
Example result
Sample input ▾
[Prospect picks up] Hello, this is Marco.
Hi Marco — this is Ava from icompaas, I'll be quick. We help SaaS teams get SOC 2-ready in weeks instead of months. Are you the right person to talk to about compliance timelines, or is that someone else on your team?
Test it on your own data