Loading…
Loading…
Compare real costs across 13 AI models — Claude, GPT-6, Gemini, DeepSeek, Grok
Workload Preset
Avg tokens in your prompt + context
Avg tokens the model generates
Total API calls across all users
Cheapest model / month
$31.50
Claude Haiku 5.5 — $0.0010/req
Most expensive model / month
$3,150
GPT-6 Astra — saves 99% vs cheapest
Total across 13 selected models / month
$11,223
30K monthly requests
Models to compare
| # | Model | / Request | Monthly |
|---|---|---|---|
| 1 | ECO Claude Haiku 5.5CHEAPEST Anthropic | $0.0010 | $31.50 |
| 2 | ECO GPT-6 Luna OpenAI | $0.0010 | $31.50 |
| 3 | ECO DeepSeek V4.1 Flash DeepSeek | $0.0025 | $76.50 |
| 4 | ECO Gemini 3.8 Flash | $0.0079 | $236.25 |
| 5 | ECO DeepSeek V4 Pro DeepSeek | $0.0086 | $257.40 |
| 6 | MID Grok 4.7 xAI | $0.013 | $390.00 |
| 7 | MID Claude Sonnet 5.5 Anthropic | $0.021 | $630.00 |
| 8 | MID GPT-6.1 Sol OpenAI | $0.021 | $630.00 |
| 9 | MID Gemini 4 Argon | $0.021 | $630.00 |
| 10 | MID Gemini 3.1 Pro Preview | $0.025 | $750.00 |
| 11 | PRO Claude Opus 5.5 Anthropic | $0.042 | $1,260 |
| 12 | PRO Claude Fable 5.1 Anthropic | $0.105 | $3,150 |
| 13 | PRO GPT-6 AstraPRICIEST OpenAI | $0.105 | $3,150 |
Highest intelligence, complex reasoning
Avg. monthly cost
$2,520
Strong performance, reasonable cost
Avg. monthly cost
$606.00
Fast, affordable for standard tasks
Avg. monthly cost
$126.63
Per 1M tokens. Sourced from official provider pricing pages.
| Model | Provider | Input $/1M | Output $/1M | Out/In ratio |
|---|---|---|---|---|
| Claude Fable 5.1 | Anthropic | $10 | $50 | 5.0x |
| GPT-6 Astra | OpenAI | $10 | $50 | 5.0x |
| Claude Opus 5.5 | Anthropic | $4 | $20 | 5.0x |
| Claude Sonnet 5.5 | Anthropic | $2 | $10 | 5.0x |
| GPT-6.1 Sol | OpenAI | $2 | $10 | 5.0x |
| Gemini 4 Argon | $2 | $10 | 5.0x | |
| Gemini 3.1 Pro Preview | $2 | $12 | 6.0x | |
| Grok 4.7 | xAI | $2 | $6 | 3.0x |
| DeepSeek V4 Pro | DeepSeek | $1.32 | $3.96 | 3.0x |
| Gemini 3.8 Flash | $0.75 | $3.75 | 5.0x | |
| DeepSeek V4.1 Flash | DeepSeek | $0.3 | $1.2 | 4.0x |
| Claude Haiku 5.5 | Anthropic | $0.1 | $0.5 | 5.0x |
| GPT-6 Luna | OpenAI | $0.1 | $0.5 | 5.0x |
Run your projects on Claude Code?
We packaged the .claude config that runs this site — 26 specialist agents, 14 workflow skills, and 6 rule files from 31 real production incidents. From $9.
Get the Claude Code Production Pack — $29AI API costs vary by 100x across providers and model tiers at October 2026 prices. A workload costing $15,000/month on Claude Fable 5.1 costs about $150 on Claude Haiku 5.5 or GPT-6 Luna at the same token volume. Understanding per-request, daily, monthly, and yearly costs across the major model families enables informed model selection — picking the cheapest model that still meets quality requirements for a given task tier.
The calculator multiplies token counts by each model's published per-million-token price for inputs and outputs separately, then sums them. Daily cost scales by requests per day, monthly by 30 days, yearly by 365. Model selection checkboxes let you narrow the comparison to relevant options. The bar chart normalises all monthly costs to the maximum to show relative scale at a glance, with cheapest and most expensive models highlighted.
A startup CTO evaluating which model to use for a customer support chatbot processing 5,000 conversations per day — the calculator shows the exact monthly delta between Sonnet 5.5 and Haiku 5.5.
A solo developer building a document-processing pipeline who needs to justify the premium of GPT-6 Astra over DeepSeek V4 Pro to their client — the yearly cost comparison makes the tradeoff concrete.
A platform team benchmarking Enterprise workloads (100K req/day) where a $0.01/request difference compounds to over $36,000/year.
A product manager choosing between Gemini 3.8 Flash and Gemini 3.1 Pro Preview — instantly sees the cost multiple for their specific input/output token mix.
Pick a preset (Light, Medium, Heavy, Enterprise) or enter your own request volume
Set average input tokens and output tokens per request using the sliders
Select which models to include in the comparison using the checkboxes
View cost per request, daily, monthly, and yearly for each model
Read the bar chart to instantly spot the cheapest and most expensive options
About the AI Model Cost Comparison Calculator
The calculator covers 13 models at the prices on each vendor pricing page (checked 13 October 2026; Gemini 4 Argon added 18 October 2026): Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, Claude Haiku 5.5 (Anthropic), GPT-6 Astra, GPT-6.1 Sol, GPT-6 Luna (OpenAI), Gemini 4 Argon, Gemini 3.1 Pro Preview, Gemini 3.8 Flash (Google), DeepSeek V4 Pro, DeepSeek V4.1 Flash, and Grok 4.7 (xAI). Short-context standard-tier rates. Gemini 4 Argon shows its introductory price, which doubles on 1 December 2026; Gemini 3.8 Flash shows its launch price, which doubles on 1 January 2027.
Cost per request = (input tokens / 1,000,000 × input price) + (output tokens / 1,000,000 × output price). Daily cost multiplies by requests per day. Monthly cost multiplies by 30. Yearly multiplies by 365.
Light (100 req/day, 500 input / 2,000 output tokens) suits small apps or personal automations. Medium (1,000 req/day) fits typical SaaS products. Heavy (10,000 req/day) covers high-traffic applications. Enterprise (100,000 req/day) represents large-scale production workloads.
Generating tokens requires more GPU computation than reading tokens. Providers typically charge 3-5x more for output. Keeping responses concise is the single most effective cost-reduction lever.
No. All calculations run entirely in your browser. No data is sent to any server. This tool is 100% client-side.
Use budget models (Haiku 5.5, GPT-6 Luna, DeepSeek V4.1 Flash) for simple tasks, reserve premium models for complex reasoning. Shorten output by asking for concise responses. Enable prompt caching for repeated system prompts. Batch requests at off-peak hours.
Your AI API bill on the day each price change lands, and the cheapest switch
Open →DeveloperFind what breaks when Chrome 158 turns XSLT off on 17 Nov 2026
Open →DeveloperHow many days your $100, $200 or $500 API credit lasts
Open →DeveloperScore your system prompt for bloat and get a cut list
Open →