Skip to content
CallMissed
Book a demo

Explore CallMissed

Back to all models
LLM Chatnewfastreasoningtools
Deepgram — deepgram-think-fast logo

Deepgram Think-Fast

by Deepgram · Released 2026

Deepgram's early-access chat model built for voice agents: follows a system prompt, calls tools and writes replies meant to be spoken. 32K context, text only.

Deepgram — deepgram-think-fast logo
LLM Chat

Deepgram Think-Fast

Powered by Deepgram · Not disclosed

Context Window

32K

Parameters

Not disclosed

Max Output

Not disclosed

Category

LLM Chat

Overview

Deepgram Think-Fast is an early-access language model built for voice agents. It follows a system prompt, calls tools, and writes replies meant to be spoken: it often says a short line such as "Sure, booking that now" alongside a tool call, so the caller hears something while the action runs.

On CallMissed you call it with the id `deepgram-think-fast` on the OpenAI-compatible `/v1/chat/completions` endpoint, or pick it as the language model of a voice agent (speech-to-text, language model, text-to-speech). It has a 32,768-token context window, supports tool calling and streaming, and takes text only.

It supports `reasoning_effort` with `none` (the default), `minimal`, `low`, `medium` and `high`; `xhigh` and `max` run as `high`, and the trace comes back as `reasoning_content`. In our tests every level above `none` took 2 to 5 seconds to answer, against about 0.3 seconds with `none`, so keep it off for voice.

Measured from our India servers, a one-sentence reply takes 0.3 to 0.8 seconds. Hindi is good; Tamil is weak. Conversations on this model are used by the provider to improve its models; the opt-out does not apply to it. It is processed outside India and is not available to India-only keys or voice agents hosted in India. As an early-access model, request fields and limits may change and there is no SLA.

Pricing

MetricPrice
Input /1M tokens₹35.0000
Output /1M tokens₹90.0000

1 credit = ₹1 ≈ US$0.0104 (US$1 = ₹96). Transparent per-call pricing — you pay only for what you call, with no per-seat charge or minimums.

Key Highlights

  • 0.3 to 0.8s for a one-sentence reply from our India servers
  • Says a short spoken line alongside a tool call
  • reasoning_effort none / minimal / low / medium / high, off by default
  • 32,768-token context window

Benchmarks

BenchmarkScore
Reply latency0.3-0.8s
With reasoning2-5s
Context32K

Technical Details

  • Context window: 32,768 tokens; text input and output only
  • Tool calling with function tools; tool_choice "auto" or "none" only
  • Streaming supported; response_format (JSON mode) is not supported
  • reasoning_effort none (default) / minimal / low / medium / high; xhigh and max run as high
  • Ignored: max_tokens, top_p, stop, seed, penalties, logprobs, n, parallel_tool_calls
  • Early access: fields and limits may change; no SLA

Strengths

  • Fast first reply with reasoning off
  • Writes replies meant to be spoken
  • Good Hindi in our tests

Limitations

  • Early access: request fields and limits may change, no SLA
  • Conversations are used by the provider to improve its models; the opt-out does not apply to it
  • 32K context and text only (no images or files)
  • Processed outside India: refused for India-only keys and hosting "india" agents
  • No JSON mode; Tamil is weak; paid plans only

Use Cases

Voice agentsPhone receptionistsTool-using call flows

API Example

curl https://api.callmissed.com/v1/chat/completions \
  -H "Authorization: Bearer cm_YOUR_KEY" \
  -d '{"model": "deepgram-think-fast", "messages": [{"role": "user", "content": "Book me a table for two at 7pm"}]}'

Endpoint: POST /v1/chat/completions · Model ID: deepgram-think-fast

Try Deepgram Think-Fast now

Get 1000 free API credits on signup. No credit card required.

Start free Book a DemoRead docs