TOKENBURN INDEX — TBX · EST. 2026
Every prompt has a burn rate.
AI is cheap until your agents start thinking out loud. TokenBurn Index measures the tokens — and the money — your AI habits quietly torch, then judges you for it.
THE BURNMETER
PASTE · MEASURE · REPENT“Your prompt works out. It just refuses to cut carbs.”
| Model | 1 run | 100 runs | 1,000 runs |
|---|---|---|---|
| Claude Fable 5 Anthropic | $0.04 | $3.60 | $36.00 |
| Claude Fable 5.1 Anthropic | $0.04 | $3.60 | $36.00 |
| Claude Opus 5 Anthropic | $0.02 | $1.80 | $18.00 |
| Claude Opus 4.8 Anthropic | $0.02 | $1.80 | $18.00 |
| Claude Opus 5.5 Anthropic | $0.01 | $1.44 | $14.40 |
| Claude Sonnet 5.5 Anthropic | $0.0072 | $0.72 | $7.20 |
| Claude Sonnet 5 Anthropic | $0.0072 | $0.72 | $7.20 |
| Claude Haiku 5.5 Anthropic | $0.0004 | $0.04 | $0.36 |
| Claude Haiku 4.5 Anthropic | $0.0036 | $0.36 | $3.60 |
| GPT-6 Astra OpenAI | $0.04 | $3.60 | $36.00 |
| GPT-5.6 OpenAI | $0.01 | $1.44 | $14.40 |
| GPT-5.6 Cyber OpenAI | $0.05 | $5.34 | $53.44 |
| GPT-5.5 OpenAI | $0.02 | $2.14 | $21.38 |
| GPT-5.4 OpenAI | $0.01 | $1.07 | $10.69 |
| GPT-5.6 Terra OpenAI | $0.0086 | $0.85 | $8.55 |
| GPT-6.1 Sol OpenAI | $0.0072 | $0.72 | $7.20 |
| GPT-6.1 Terra OpenAI | $0.0036 | $0.36 | $3.60 |
| GPT-6 Sol OpenAI | $0.0072 | $0.72 | $7.20 |
| GPT-5 Mini OpenAI | $0.0014 | $0.14 | $1.41 |
| GPT-5.6 Luna OpenAI | $0.0009 | $0.09 | $0.85 |
| GPT-6 Luna OpenAI | $0.0004 | $0.04 | $0.36 |
| Gemini 4 Argon Google | $0.0072 | $0.72 | $7.20 |
| Gemini 3.1 Pro Google | $0.0086 | $0.85 | $8.55 |
| Gemini 3.6 Flash Google | $0.0027 | $0.27 | $2.70 |
| Gemini 3.5 Flash Google | $0.0064 | $0.64 | $6.41 |
| Gemini Omni Flash Google | $0.0064 | $0.64 | $6.41 |
| Gemini 3.7 Flash Google | $0.0027 | $0.27 | $2.70 |
| Gemini 3.8 Flash Google | $0.0027 | $0.27 | $2.70 |
| Gemini 3 Flash Google | $0.0021 | $0.21 | $2.14 |
| Gemini 3.5 Flash Lite Google | $0.0018 | $0.18 | $1.76 |
| DeepSeek V4 Pro DeepSeek | $0.0015 | $0.15 | $1.49 |
| DeepSeek V4.1 Flash DeepSeek | $0.0004 | $0.04 | $0.44 |
| GLM 5.3 Z.ai | $0.0033 | $0.33 | $3.29 |
| GLM 5.3 Flash Z.ai | $0.0004 | $0.04 | $0.37 |
| Kimi K2.6 Moonshot AI | $0.0029 | $0.29 | $2.91 |
| Kimi K3.1 Pro Moonshot AI | $0.0072 | $0.72 | $7.20 |
| Kimi K3 Moonshot AI | $0.01 | $1.08 | $10.80 |
| Muse Wonder Meta | $0.01 | $1.44 | $14.40 |
| Muse Spark 1.3 Meta | $0.0032 | $0.32 | $3.15 |
| Grok 4.7 xAI | $0.0045 | $0.45 | $4.50 |
| MiMo V2.6 Flash Xiaomi | $0.0002 | $0.02 | $0.22 |
you could help me with something. the tone should be friendly but professional. make sure the tone is friendly but professional at all times. ---------------------------------------- **THE TASK** write a short email to my team letting them know that the meeting moved from 3pm to 4pm? keep it short. keep it short, and make sure the tone is friendly but professional. The meeting moved from 3pm to 4pm. the email should be short. a simple, short email. the time change. ----------------------------------------
Count the damage
Paste a prompt, a conversation, an agent workflow. Get estimated tokens, cost across 41 models, and a Token Burn Score from 0 to 100.
Same text, 208x price range
The same request costs $0.0004 or $0.09 depending on the model. Most teams never look. You are about to.
Get judged, then optimized
A dry-humored diagnosis, a leaner version of your prompt, and the savings at 1,000 runs. Shame is a feature, not a bug.
npx tokenburnindex auditReads the logs Claude Code, Codex and opencode already keep on your disk and shows where the money went. Nothing leaves your machine.
AI WASTE HABITS — 2026
ALL HABITS →Re-pasting the entire conversation "for context"
The model already had the context. Now it has it twice, at input prices.
Agents left on auto-approve overnight
You woke up to 340 tool calls and a refactor nobody asked for.
A 2,000-token system prompt for a yes/no question
The answer was 'no'. The invoice was not.
COST TRANSPARENCY INDEX
FULL INDEX →10 AI providers scored against 25 yes/no criteria — can a customer know what they will be charged, and see where the money went? Every score cites dated primary evidence.
HALL OF SHAME
ALL CASES →“I used 40,000 tokens to ask for a button color.”
EST. 40,000 TOKENS · $0.42 · The button is still blue.
“My agent rewrote half the repo to rename a variable.”
EST. 1,200,000 TOKENS · $14.70 · The variable is now camelCase. The budget is not.
$AI is cheap until your agents start thinking out loud.
$More tokens. Same result. Beautifully expensive.
$Your AI workflow is not free. It is just quiet.
$Your agent did not work harder. It just spent more.
AI Tokenomics, explained
What tokens actually are, why output costs 4-8x more than input, what cached tokens change, and why agents multiply everything. No hype, no doom — just the meter.