> ## Documentation Index
>
> Fetch the complete documentation index at: https://lithi.ai/llms.txt
>
> Use this file to discover all available pages before exploring further.

# Compare the cost of your AI work.

Compare Lithi Compute with model APIs, AWS and Google Cloud using published workload examples. Check the unit, included work and assumptions.

- **Model APIs:** Compare the same token counts, with each model's published rule for media.
- **Cloud services:** Provider-specific units, allowances and included work.
- **Complete workflow:** The steps included in each total, from input to output.

[Get a free estimate](/compute) · [Pricing and assumptions](/pricing)

## Compare model API prices

The model API columns use published batch rates. Each row states its workload quantity and any model-specific media-token rule.

| Job | Lithi | [gpt-6-luna (Batch)](https://developers.openai.com/api/docs/pricing) | [Gemini 3.1 Flash-Lite (Batch)](https://ai.google.dev/gemini-api/docs/pricing) | [Claude Haiku 4.5 (Batch)](https://platform.claude.com/docs/en/about-claude/pricing) |
| --- | --- | --- | --- | --- |
| **Documents** |  |  |  |  |
| Translate 5,000 documents (1,500 words each, into 10 languages) | $15.25 | $30.25 | $88.13 | $302.50 |
| Read 100,000 scanned pages (1,024-pixel page image in, about 500 tokens of text out) | $25.98 | No published rate: OpenAI has not published an image-token rule for this model. | $51.65 (own media-token rule) | $198.45 (own media-token rule) |
| **Code, for your coding agent** |  |  |  |  |
| Review 10,000 code changes (8,000 tokens in, a 1,000-token report out) | $5.00 | $6.50 | $17.50 | $65.00 |
| Debug 5,000 failing tests (9,600 tokens in, a 1,200-token report out) | $3.00 | $3.90 | $10.50 | $39.00 |
| Map 2,000 modules for your agent (9,600 tokens in, a 1,200-token summary out) | $1.20 | $1.56 | $4.20 | $15.60 |
| **Data and evaluation** |  |  |  |  |
| Label 1,000,000 items (9,600 tokens in, 1,200 out per item) | $600.00 | $780.00 | $2,100.00 | $7,800.00 |
| Classify 5,000,000 records (200 tokens in, 10 out per record) | $55.00 | $62.50 | $162.50 | $625.00 |
| Score 1,000,000 answers (2,000 tokens in, 100 out per answer) | $110.00 | $125.00 | $325.00 | $1,250.00 |
| **Images** |  |  |  |  |
| Describe 5,000 images (1,024 × 1,024 pixels, one pass each) | $1.30 | No published rate: OpenAI has not published an image-token rule for this model. | $2.58 (own media-token rule) | $9.92 (own media-token rule) |
| **Video: 500 hours into 10 languages** |  |  |  |  |
| Transcribe the speech (30,000 minutes) | $2.85 | No published rate: The provider publishes no audio input rate for this model. | $18.90 (own media-token rule) | No published rate: The provider publishes no audio input rate for this model. |
| Translate the subtitles (6,000,000 tokens into each of 10 languages) | $9.15 | $18.15 | $52.88 | $181.50 |
| Find scenes and keyframes (180,000 scenes, one about every 10 seconds) | $15.75 | No published rate: The provider publishes no video input rate for this model. | $63.23 (own media-token rule) | No published rate: The provider publishes no video input rate for this model. |
| Choose the clips (1,000 calls, one per 30 minutes of video) | $0.60 | $0.78 | $2.10 | $7.80 |
| All four video steps (1 + 2 + 3 + 4) | $28.35 | No published rate: gpt-6-luna and Claude Haiku 4.5 publish no audio or video input rate, so they have no total for all four steps. | $137.11 (own media-token rule) | No published rate: gpt-6-luna and Claude Haiku 4.5 publish no audio or video input rate, so they have no total for all four steps. |

¹ Counted with that model's own media-token rule. — The model publishes no rate for this kind of input.

Each model counts tokens its own way. Lithi figures use the published example metering profile; a signed quote sets the real charge. These are prices for the same work, not a test of output quality. Prices checked 5 October 2026.

Lithi charges $0.05 per 1M input tokens and $0.10 per 1M output tokens. In the 9 jobs where every model publishes a rate on the same token counts, the Lithi total is lowest in 9. Sources checked 5 October 2026. Published list prices from each provider's pricing page; the full working is on the [pricing page](/pricing).

## Compare cloud-service costs

Cloud services bill per page, minute, character or image. Each row keeps its native unit, allowance and included work beside the total.

| Job | Lithi Compute | AWS | Google Cloud |
| --- | --- | --- | --- |
| Translate 5,000 documents (1,500 words each, into 10 languages) | $15.25 | $6,187.50 — 412,500,000 characters ([Amazon Translate](https://aws.amazon.com/translate/pricing/). Free monthly allowances are ignored at this volume.) | $8,250.00 — 412,500,000 characters ([Google Cloud Translation](https://cloud.google.com/translate/pricing). Free monthly allowances are ignored at this volume.) |
| Read 100,000 scanned pages (1,024-pixel page image in, about 500 tokens of text out) | $25.98 | $150.00 — 100,000 pages ([Amazon Textract DetectDocumentText](https://aws.amazon.com/textract/pricing/)) | $150.00 — 100,000 pages ([Google Document AI Enterprise OCR](https://cloud.google.com/document-ai/pricing). The first 1,000 free pages each month are ignored at this volume.) |
| Label 1,000,000 items (9,600 tokens in, 1,200 out per item) | $600.00 | $42,000.00 — 1,000,000 objects ([Amazon SageMaker Ground Truth automated labelling](https://aws.amazon.com/sagemaker/ai/pricing/). An automated labelling service with its own workflow, not an AI-model call.) | No published rate (Google Cloud data labeling. Google Cloud publishes no current price for an automated labelling service.) |
| Describe 5,000 images (1,024 × 1,024 pixels, one pass each) | $1.30 | $5.00 — 5,000 images ([Amazon Rekognition Image label detection](https://aws.amazon.com/rekognition/pricing/)) | $7.50 — 5,000 units ([Google Cloud Vision label detection](https://cloud.google.com/vision/pricing). The first 1,000 free units each month are ignored at this volume.) |
| Transcribe the speech (30,000 minutes) | $2.85 | $180.00 — 30,000 minutes ([Amazon Transcribe (batch)](https://aws.amazon.com/transcribe/pricing/)) | $480.00 — 30,000 minutes ([Google Speech-to-Text V2 (standard)](https://cloud.google.com/speech-to-text/pricing))<br>$90.00 — 30,000 minutes ([Google Speech-to-Text V2 (dynamic batch)](https://cloud.google.com/speech-to-text/pricing)) |
| Translate the subtitles (6,000,000 tokens into each of 10 languages) | $9.15 | $3,712.50 — 247,500,000 characters ([Amazon Translate](https://aws.amazon.com/translate/pricing/). Free monthly allowances are ignored at this volume.) | $4,950.00 — 247,500,000 characters ([Google Cloud Translation](https://cloud.google.com/translate/pricing). Free monthly allowances are ignored at this volume.) |
| Find scenes and keyframes (180,000 scenes, one about every 10 seconds) | $15.75 | $1,500.00 — 30,000 minutes ([Amazon Rekognition Video shot detection](https://aws.amazon.com/rekognition/pricing/))<br>$3,000.00 — 30,000 minutes ([Amazon Rekognition Video label detection](https://aws.amazon.com/rekognition/pricing/)) | $1,500.00 — 30,000 minutes ([Google Video Intelligence shot detection](https://cloud.google.com/video-intelligence/pricing). The free first minutes each month are ignored at this volume.)<br>$3,000.00 — 30,000 minutes ([Google Video Intelligence label detection](https://cloud.google.com/video-intelligence/pricing). The free first minutes each month are ignored at this volume.) |
| Choose the clips (1,000 calls, one per 30 minutes of video) | $0.60 | No published rate (Managed clip selection. AWS and Google Cloud publish no price for a managed clip-selection service.) | No published rate (Managed clip selection. AWS and Google Cloud publish no price for a managed clip-selection service.) |

Each service bills in its own unit, and a managed service is not an AI-model call. These are prices for the same work, not a test of output quality. Sources checked 5 October 2026. Published list prices from each provider's pricing page; the full working is on the [pricing page](/pricing).

## What the video workflow costs

This example covers 500 hours of video and 10 languages. Read each provider's covered steps before comparing its amount.

Assumptions: 500 hours of video; 10 languages; Lithi covers transcription, translation, scene finding and clip selection. Step order: 1 + 2 + 3 + 4.

| Provider | Transcribe | Translate | Scenes | Clips | Amount |
| --- | --- | --- | --- | --- | --- |
| Lithi Compute | Yes | Yes | Yes | Yes | $28.35 — 4-step total |
| Google | Yes | Yes | Yes | Yes | $137.11 — 4-step total |
| AWS | Yes | Yes | Yes | Not included in this total | $5,392.50 — 3-step subtotal |
| Google Cloud | Yes | Yes | Yes | Not included in this total | $6,540.00 — 3-step subtotal |

<details>
<summary>Per-step prices, products and sources</summary>

Lithi Compute: $21.09 input-token cost + $7.26 output-token cost for the complete workflow.

### Google

- Transcribe the speech: $18.90 — [Gemini 3.1 Flash-Lite (Batch)](https://ai.google.dev/gemini-api/docs/pricing)
- Translate the subtitles: $52.88 — [Gemini 3.1 Flash-Lite (Batch)](https://ai.google.dev/gemini-api/docs/pricing)
- Find scenes and keyframes: $63.23 — [Gemini 3.1 Flash-Lite (Batch)](https://ai.google.dev/gemini-api/docs/pricing)
- Choose the clips: $2.10 — [Gemini 3.1 Flash-Lite (Batch)](https://ai.google.dev/gemini-api/docs/pricing)

### AWS

- Transcribe the speech: $180.00 — [Amazon Transcribe (batch)](https://aws.amazon.com/transcribe/pricing/)
- Translate the subtitles: $3,712.50 — [Amazon Translate](https://aws.amazon.com/translate/pricing/)
- Find scenes and keyframes: $1,500.00 — [Amazon Rekognition Video shot detection](https://aws.amazon.com/rekognition/pricing/)

### Google Cloud

- Transcribe the speech: $90.00 — [Google Speech-to-Text V2 (dynamic batch)](https://cloud.google.com/speech-to-text/pricing)
- Translate the subtitles: $4,950.00 — [Google Cloud Translation](https://cloud.google.com/translate/pricing)
- Find scenes and keyframes: $1,500.00 — [Google Video Intelligence shot detection](https://cloud.google.com/video-intelligence/pricing)

</details>

A dash marks a step excluded from the displayed amount. AWS and Google Cloud subtotals omit clip selection; this is a comparison of included work, not every capability a provider offers. gpt-6-luna and Claude Haiku 4.5 publish no audio or video input rate, so they have no total for all four steps. Sources checked 5 October 2026. Published list prices from each provider's pricing page; the full working is on the [pricing page](/pricing).

## Choose by workflow fit

The right choice depends on the work you need covered, the unit you want to pay in and the service around it.

### Lithi Compute may suit you if

- You have batch work your AI agent can delegate, sized in tokens: documents, scans, transcripts, code review or labelling.
- You want one tariff across jobs instead of a different unit for every service.
- You want a free estimate for your own counts before any work starts.

### A cloud service may suit you if

- You need a managed service inside your existing AWS or Google Cloud account, contract or region.
- You rely on a feature of that service beyond the price of the job, such as its own workflow or tooling.
- Your work is outside the jobs Lithi Compute supports.

### Before comparing totals

- Lithi runs supported jobs only. Check that your job is on the supported work list before you plan around the price.
- The Lithi figures use the published example metering profile. A signed quote sets the real charge.
- Model APIs count tokens their own way, so the same text can be a different number of tokens.

## Questions before you choose

### When is Lithi a good fit? {#good-fit}

Start with a bounded task whose input you can prepare and whose result you can check. Repeatable code, data, document and media work are useful starting points. Compare the required quality, preparation, review and delivery before comparing cost.

### Does Lithi replace AWS or Google Cloud? {#cloud-services}

Lithi handles suitable AI work through your existing agent. Cloud services also provide broader application and infrastructure capabilities. Use Lithi where its supported tasks, data requirements and output checks fit your work.

### Can Lithi cost less for my workload? {#cost-comparison}

It depends on the job, so the tables show each one. AWS and Google Cloud services bill in their own units for their own kind of service, so those rows set published prices side by side without a ranking. Against AI model APIs counted on the same tokens, the Lithi total is lowest in 9 of the 9 jobs where every model publishes a rate.

### What does Lithi charge? {#compute-charges}

One tariff: $0.05 per 1M input tokens and $0.10 per 1M output tokens. Media converts to tokens through a signed metering recipe, and a signed quote sets the real charge for your work.

### How current are these prices? {#price-freshness}

Every price was checked on 5 October 2026 against the provider's own pricing page, linked beside the figure. Providers change prices, so confirm before you commit. The pricing page shows the full working for each job.

### Do the totals include free allowances? {#free-allowances}

No. Where a service has a free monthly allowance, the note beside the figure says it is ignored at this volume, so the total is the list price for the whole job.

### Does this compare output quality? {#output-quality}

No. The tables compare published prices for the same work, not the quality of the output. Each model also counts tokens its own way, so the same text can be a different number of tokens.



Use your own workload counts to get a free estimate, then read the assumptions behind these published examples.

[Get a free estimate](/compute)

- [Pricing and assumptions](/pricing)
- [Compute documentation](/docs/compute)