Accuracy is a cost decision, not a quality decision
A model with 12% word error rate versus one with 4% doesn't just read worse — it converts directly into human review minutes. At $28/hour, six extra review minutes per file costs $2.80, which is 10× what you'd save picking the cheaper API on a 45-minute recording. Always price the review line before optimising the API line.
Where the add-ons hide
Diarization (who spoke when), timestamps, translation, and summarization are often billed separately or require a second model pass. Summarization in particular means sending the entire transcript back through an LLM — a 45-minute recording is roughly 9,000 input tokens every time.
Batch and async pricing
Most providers discount non-real-time transcription substantially versus streaming. If your workflow can tolerate a few minutes of latency, batch endpoints commonly cut per-minute cost by 30–50%. Reserve real-time pricing for genuinely live use cases.
Related guides
Long-form playbooks on the same topic, written by the RevenueLab editorial team.
LLM Token Costs in 2026: Pricing Every Model, Hidden Multipliers, and Margin Math
Input vs output token pricing across GPT, Claude, and Gemini, the context-window cost trap, how caching and batching cut bills 40–80%, and the real per-user margin most AI apps miss.
Read the guideSaaS Pricing Strategy: Per-Seat, Usage, Tiers, and the Hybrid Future
A framework for choosing a SaaS pricing model — when per-seat caps your growth, when usage-based makes revenue volatile, and how hybrid models stitch the two together.
Read the guideFAQ
How much does AI transcription cost per hour?
API-only, expect roughly $0.18–$1.20 per audio hour at $0.003–$0.02 per minute. Add summarization and the total lands around $0.30–$2.00 per hour. Add human review at $28/hour and a lightly-reviewed hour costs $3–$6 all-in.
Is AI transcription cheaper than human transcription?
Dramatically. Human transcription runs $60–$180 per audio hour. AI plus a light human review pass typically lands at $3–$8 per hour — a 90%+ reduction, at accuracy that's acceptable for most non-legal use cases.
What's a good word error rate?
3–8% on clean single-speaker audio is strong. Multi-speaker, accented, or noisy audio commonly runs 10–20%, and that's where review costs explode. Test on your actual audio, not a demo clip.
Does diarization cost extra?
Often yes — either as a price uplift on the transcription call or as a separate processing step. Check whether your provider bundles it before assuming the headline per-minute rate is your true cost.
How many tokens is an hour of audio?
Roughly 12,000 tokens (150 words/minute × 60 minutes × ~1.33 tokens/word). That's what you resend to an LLM every time you summarize or ask questions about the transcript.
Can I cut cost by transcribing only part of the audio?
Yes — voice activity detection to strip silence, and trimming intros/outros, can cut billable minutes 10–25% on recordings with dead air. Meeting recordings benefit the most.
How this calculator is built
Independently maintained
Written by Sam Doshi and the RevenueLab editorial team. We don't sell the data feeds this tool is built on.
Sourced from primary data
Benchmarks come from public AdSense / Stripe / IRS disclosures and reader-submitted data — never third-party "$X per view" claims. Full methodology.
Last editorial review
Reviewed on a rolling quarterly cycle. Dated reviews are published on the methodology record for each calculator.
Editorial standards
See our editorial policy and disclaimer. Results are estimates, not advice.