Best cheap speech-to-text API
4 models · updated 2026-07-15
The verdict
Groq leads — 1 of 4 models rank Groq the top pick.
Not unanimous: ChatGPT picks Cloudflare Workers AI; Gemini picks DeepInfra; Grok picks Rev AI.
As of 2026-07-15, ChatGPT, Claude, Gemini, Grok collectively rank Groq first for cheap speech-to-text api on modelsagree.com.
Your vendor missing? Check any brand →
Combined ranking
- 1GPT #2Claude #1Gemini #3Grok —
Cheapest hosted transcription that's still genuinely accurate — roughly $0.04/hour (~$0.0007/min) for OpenAI's Whisper large-v3-turbo weights served on Groq's LPUs, with faster-than-real-time turnaround measured in seconds; for the typical developer batch-transcribing podcasts, calls, or user audio, nothing hosted comes close on price-per-accurate-minute
+ model takes & fixes− hide details
Claude Cheapest hosted transcription that's still genuinely accurate — roughly $0.04/hour (~$0.0007/min) for OpenAI's Whisper large-v3-turbo weights served on Groq's LPUs, with faster-than-real-time turnaround measured in seconds; for the typical developer batch-transcribing podcasts, calls, or user audio, nothing hosted comes close on price-per-accurate-minute
GPT Near-tied for first: accurate multilingual transcription at $0.04/hour, extremely fast inference, and an easy OpenAI-compatible API make it unusually practical.
Gemini Delivers near-instantaneous transcription speed (up to 200x real-time) at $0.00067/min for Whisper Large V3 Turbo, using highly optimized LPU hardware.
Where it falls shortper GPT File transcription rather than true streaming, with minimum-duration billing and fewer speech-specific features than specialist APIs.
per Claude Batch/pre-recorded only with a bare-bones feature set — no streaming, weak diarization, and rate limits on lower tiers make it wrong for real-time voice agents or enterprise SLAs
per Gemini Subject to strict rate limits on request concurrency and volume, making it difficult to scale without enterprise negotiation.
- 2GPT #1Claude —Gemini #2Grok —
Exceptional batch value at $0.0005/audio minute, with strong multilingual Whisper accuracy and a generous daily free allocation; narrowly beats Groq on raw cost.
+ model takes & fixes− hide details
GPT Exceptional batch value at $0.0005/audio minute, with strong multilingual Whisper accuracy and a generous daily free allocation; narrowly beats Groq on raw cost.
Gemini Extremely cost-effective at $0.00045/min for Whisper and $0.00051/min for Whisper Large V3 Turbo, with edge-hosted serverless execution and a generous free daily neuron tier.
Where it falls shortper GPT Batch-oriented and comparatively bare-bones—no native diarization or polished transcription workflow, and sustained usage requires a paid Workers plan.
per Gemini Bound by strict file size and execution limits inherent to serverless workers, requiring audio files to be pre-split into small segments.
- 3GPT #4Claude #5Gemini #4Grok —
At $0.21/hour, it pairs strong messy-speech and entity accuracy with prompting, custom spelling, language detection, formatting, and word timestamps; worth the small premium for transcript usability.
+ model takes & fixes− hide details
GPT At $0.21/hour, it pairs strong messy-speech and entity accuracy with prompting, custom spelling, language detection, formatting, and word timestamps; worth the small premium for transcript usability.
Gemini Exceptional accuracy on noisy audio and accents for $0.0025/min (Universal-2) or $0.0035/min (Universal-3 Pro Async), combined with robust speaker diarization and audio intelligence tools.
Claude Aggressive price cuts brought Universal to ~$0.0025-0.003/min (~$0.15/hr) with accuracy competitive with Deepgram and the strongest bundled extras at this price — diarization, sentiment, PII redaction, and audio-intelligence add-ons that would cost extra elsewhere
Where it falls shortper GPT Supports far fewer languages than Whisper or Speechmatics, and diarization or other intelligence features can raise the effective price.
per Claude Cheapest rates assume prepaid/volume tiers and the add-ons that justify choosing it each bill separately; pure transcription users on small volumes pay more per minute than Groq or gpt-4o-mini-transcribe
per Gemini Advanced features (like diarization and summarization) carry modular add-on costs that can quickly double or triple the base price.
- 4GPT —Claude #2Gemini #5Grok —
Best accuracy-per-dollar among purpose-built STT vendors at ~$0.0043/min pay-as-you-go; Nova-3 beats Whisper-class models on noisy, multi-speaker, real-world audio, and you get streaming, diarization, keyterm prompting, and smart formatting in one API — the pick when the audio is hard, not just the price low
+ model takes & fixes− hide details
Claude Best accuracy-per-dollar among purpose-built STT vendors at ~$0.0043/min pay-as-you-go; Nova-3 beats Whisper-class models on noisy, multi-speaker, real-world audio, and you get streaming, diarization, keyterm prompting, and smart formatting in one API — the pick when the audio is hard, not just the price low
Gemini Optimized for real-time streaming conversational applications with pricing starting around $0.0043/min to $0.0052/min for batch, utilizing highly efficient Nova-2 and Nova-3 models.
Where it falls shortper Claude 5-6x Groq's price, so pure-batch users with clean audio overpay; volume discounts require committed-spend contracts
per Gemini Streaming endpoints are more expensive ($0.0077/min) and all intelligence features must be paid for as separate add-ons.
- 5GPT —Claude —Gemini #1Grok —
Offers the lowest raw pricing on the market at $0.00020/min for Whisper Large V3 Turbo and $0.00045/min for Whisper Large V3, running on high-concurrency serverless GPU infrastructure.
+ model takes & fixes− hide details
Gemini Offers the lowest raw pricing on the market at $0.00020/min for Whisper Large V3 Turbo and $0.00045/min for Whisper Large V3, running on high-concurrency serverless GPU infrastructure.
Where it falls shortper Gemini Provides only raw transcription outputs without advanced features like diarization, custom vocabulary, or audio intelligence.
- 6GPT —Claude —Gemini —Grok #1
Lowest pricing at ~$0.002/min ($0.12/hr) for English batch/async with solid real-world accuracy for cost-sensitive use cases, strong on long-form/archival audio per benchmarks; ideal for typical practitioners needing high volume without
+ model takes & fixes− hide details
Grok Lowest pricing at ~$0.002/min ($0.12/hr) for English batch/async with solid real-world accuracy for cost-sensitive use cases, strong on long-form/archival audio per benchmarks; ideal for typical practitioners needing high volume without
- 7GPT —Claude #3Gemini —Grok —
$0.003/min with word error rates better than the original whisper-1 endpoint ($0.006/min), trivially adopted since most stacks already have an OpenAI key; near-tie with Deepgram — it wins on simplicity and raw WER on clean audio, Deepgram wins on features
+ model takes & fixes− hide details
Claude $0.003/min with word error rates better than the original whisper-1 endpoint ($0.006/min), trivially adopted since most stacks already have an OpenAI key; near-tie with Deepgram — it wins on simplicity and raw WER on clean audio, Deepgram wins on features
Where it falls shortper Claude No speaker diarization and limited timestamp granularity, so it's not for meeting-transcription products; as an LLM-based transcriber it can occasionally hallucinate fluent-but-wrong text on silence or noise
- 8GPT #3Claude —Gemini —Grok —
$0.129/hour with broad accent and 56+ language coverage, per-second billing, diarization, timestamps, and formatting—arguably the best cheap full-featured transcription API.
+ model takes & fixes− hide details
GPT $0.129/hour with broad accent and 56+ language coverage, per-second billing, diarization, timestamps, and formatting—arguably the best cheap full-featured transcription API.
Where it falls shortper GPT Batch-only at this price; enhanced accuracy and real-time operation cost materially more.
- 9GPT —Claude #4Gemini —Grok —
Zero per-minute cost — the CTranslate2 reimplementation runs Whisper large-v3 ~4x faster than reference on a single consumer or spot-priced GPU, so effective cost lands around $0.0001-0.0005/min at even modest volume, with full data privacy; at sustained scale this beats every commercial API
+ model takes & fixes− hide details
Claude Zero per-minute cost — the CTranslate2 reimplementation runs Whisper large-v3 ~4x faster than reference on a single consumer or spot-priced GPU, so effective cost lands around $0.0001-0.0005/min at even modest volume, with full data privacy; at sustained scale this beats every commercial API
Where it falls shortper Claude You operate the infrastructure — GPU provisioning, scaling, and no support; below roughly a few hundred hours/month the engineering time costs more than just paying Groq
- 10GPT #5Claude —Gemini —Grok —
Strong multilingual accuracy and mature cloud controls for $0.003/minute in asynchronous dynamic-batch mode, especially attractive for large archives already on Google Cloud.
+ model takes & fixes− hide details
GPT Strong multilingual accuracy and mature cloud controls for $0.003/minute in asynchronous dynamic-batch mode, especially attractive for large archives already on Google Cloud.
Where it falls shortper GPT The bargain price requires delay-tolerant batch processing; ordinary real-time or synchronous recognition is far more expensive.
Just missed the top 5
GPT Deepgram Nova-3 — excellent noisy-audio and streaming performance, but its $0.0048–$0.0077/minute pricing trails stronger budget values · faster-whisper — can be extremely cheap at sustained self-hosted volume, but infrastructure, scaling, diarization, and idle-capacity costs prevent a universally credible per-minute API price
Claude Speechmatics — excellent multilingual accuracy but ~$0.008/min+ pricing puts it outside the "cheap" bracket · Fireworks AI Whisper serving — comparably cheap Whisper-as-a-service to Groq, but slower, less proven at scale, and adds nothing over Groq's offering
Gemini OpenAI Whisper API — priced at $0.006/min, which is up to 30x more expensive than DeepInfra and has slower processing times · Gladia — priced at $0.010/min, which makes it too expensive to compete on a budget despite its high-quality bundled features
By model
ChatGPT
- 1.Cloudflare Workers AI
- 2.Groq
- 3.Speechmatics
- 4.AssemblyAI
- 5.Google Cloud Speech-to-Text
Claude
- 1.Groq
- 2.Deepgram
- 3.OpenAI gpt-4o-mini-transcribe
- 4.faster-whisper
- 5.AssemblyAI
Gemini
- 1.DeepInfra
- 2.Cloudflare Workers AI
- 3.Groq
- 4.AssemblyAI
- 5.Deepgram
Grok
- 1.Rev AI
Common questions
What is the best cheap speech-to-text api according to AI models?
Groq leads. 1 of 4 models rank Groq the top pick. The current top 3: Groq, Cloudflare Workers AI, AssemblyAI. Ranked by asking ChatGPT, Claude, Gemini, Grok the same buying question and merging their top-5 picks, updated 2026-07-15. Source: modelsagree.com.
Which cheap speech-to-text api did each AI model pick first?
ChatGPT: Cloudflare Workers AI. Claude: Groq. Gemini: DeepInfra. Grok: Rev AI.
Do the AI models agree on the best cheap speech-to-text api?
Not unanimous. ChatGPT picks Cloudflare Workers AI; Gemini picks DeepInfra; Grok picks Rev AI.
How is this cheap speech-to-text api ranking made?
ChatGPT, Claude, Gemini, Grok are each asked the same buying question in a fresh session with no system steering. Their top-5 answers are merged (rank 1 = 5 pts … rank 5 = 1 pt) into the consensus ranking, re-polled weekly and tracked over time.
More on how polling works: full methodology →
This ranking moves
We re-poll all four models weekly. Get one short email when a #1 flips.
Cite this ranking
ModelsAgree, “Best cheap speech-to-text API” — merged ranking from ChatGPT, Claude, Gemini & Grok, polled 2026-07-15. https://modelsagree.com/best/best-cheap-speech-to-text-api (CC BY 4.0)
Tracked by ModelsAgree · rank 1 = 5 pts … rank 5 = 1 pt · re-polled weekly