CONVERSATION → TOKENS

How many tokens is a ChatGPT conversation?

Last updated

A typical 20-turn conversation with ~150 words per turn contains about 4,000 tokens of visible text — but because the model re-reads the whole history on every turn, the cumulative billed input across those 20 turns is several times higher.

QUICK CONVERTER

≈ Words

3,000

≈ Tokens

3,990

Estimates use ~1.33 tokens per English word. For an exact count of real text — including code, JSON, and other languages — paste it into the token counterand choose your model.

The detail

Conversations are the workload where token math surprises people most. Chat models are stateless: each new message sends the system prompt plus the entire visible history again. The 20th turn of a conversation bills for all 19 previous turns as input, so cumulative input tokens grow roughly quadratically with conversation length.

Concretely: 20 turns averaging 200 tokens each hold 4,000 tokens of text, but the cumulative input billed across the conversation approaches 42,000 tokens before caching. This is why long support chats and agent loops cost far more than their transcripts suggest, and why prompt caching (which discounts the unchanged prefix, typically by ~90%) matters so much for chat products.

The same effect eventually hits context limits: apps silently drop or summarize the oldest turns once the window fills. If a long conversation "forgets" early details, it likely scrolled out of context — token-count your history to know when.

Conversation sizes (at ~150 words per turn)

Turn~Words~Tokens
5-turn chat750998
10-turn chat1,5001,995
20-turn chat3,0003,990
50-turn chat7,5009,975
100-turn chat15,00019,950

FAQ

How many tokens is a ChatGPT conversation? — FAQ

Why does a long conversation cost more per message?

Because every message re-sends the full history as input. Turn 50 bills for 49 prior turns plus the system prompt, so per-message cost grows with conversation length even when your messages stay short.

How do I keep conversation token costs down?

Summarize old turns into a compact recap, start fresh threads for new topics, and use prompt caching so the unchanged history prefix bills at the cached rate (typically ~90% off).

How many tokens before ChatGPT forgets the start of a chat?

When history plus your new message exceeds the model context window, the oldest turns are dropped or compressed. Count your conversation with the chat token counter to see how close you are.