Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

Output tokens

Output tokens are the tokens a language model generates in its reply. They are counted and billed separately from input tokens, and a max_tokens setting caps how many one response may contain.

Also called: max output tokens, max_tokens, completion tokens

By Divya Sharma · Checked Nov 15, 2026 · How we review

Every call to a language model has two token counts. Input tokens cover everything you send: the system prompt, history, documents and tool definitions. Output tokens are what the model writes back. API responses usually report both, and providers price them separately; see token pricing and the site's AI model pricing research.

max_tokens

Most APIs take a parameter, often called max_tokens, that sets an upper limit on output length. Anthropic's Messages API describes it as the maximum number of tokens to generate before stopping, and notes that a model may stop earlier. If a reply hits the cap, it is cut off, and the response reports why it stopped. Each model also has its own ceiling for this value, separate from its context window.

Why it matters

- Cost: long answers cost more, and on many price lists output tokens cost more than input tokens. - Speed: output is generated one token at a time, so length drives total response time. See tokens per second. - Truncation: a cap that is too low breaks JSON and code halfway through.

Reasoning tokens

Reasoning models and chain-of-thought prompts produce extra output for the thinking steps. Depending on the provider, those tokens may count toward the cap and the bill even when they are not shown to the user.

Practical tips

Ask for concise answers when you do not need long ones, set max_tokens with headroom for the longest valid reply, and check the stop reason in code so truncated replies are retried or flagged rather than silently used.

Related terms

Token (LLM)
A token (LLM) is the basic unit of text a language model reads and writes, often a word, part of a word, a punctuation mark or a byte. Prices, speed and context limits are all counted in tokens.
Token pricing
Token pricing is how most AI model APIs charge: a set price per million input tokens you send and a separate, usually higher, price per million output tokens the model generates, with discounts for cached input and batch jobs.
Context window
A context window is the maximum amount of text, counted in tokens, that a language model can consider at once, covering the system prompt, conversation, documents, tool results and the response it writes.
Tokens per second
Tokens per second (tok/s) measures how fast a language model produces output. It is quoted either per user, the speed one reply streams at, or per system, the total output across all concurrent requests.
Reasoning model
A reasoning model is a language model trained to work through a problem in a stream of intermediate thinking tokens before giving its final answer, which tends to help on maths, coding and multi-step tasks.

Sources