Models that can take a whole document
The 27 of 60 models Else can run with a context window of a million tokens or more — enough for a long document and a long answer in one request.
A model’s context window is the ceiling on everything in one request at once: your instructions, the material you attach, and the answer being written. 27 of the 60 models Else can run have a window of 1M tokens or more, which is enough to put a long document in and still have room for a long answer.
The models
Section titled “The models”| Model | Runs as | Context | Input / 1M tokens | Output / 1M tokens |
|---|---|---|---|---|
gemini-2.5-flash-lite |
Exact engine pick | 1M | $0.12 | $0.48 |
gpt-4.1-nano |
Exact engine pick | 1M | $0.12 | $0.48 |
gemini-3.1-flash-lite |
Exact engine pick | 1M | $0.30 | $1.80 |
gemini-2.5-flash |
Exact engine pick | 1M | $0.36 | $3.00 |
gemini-3.5-flash-lite |
Exact engine pick | 1M | $0.36 | $3.00 |
m3 |
Exact engine pick | 1M | $0.36 | $1.44 |
gpt-4.1-mini |
Exact engine pick | 1M | $0.48 | $1.92 |
gemini-3-flash |
Exact engine pick | 1M | $0.60 | $3.60 |
gpt-5.6-luna |
Exact engine pick | 1.1M | $1.20 | $7.20 |
gemini-2.5-pro |
Exact engine pick | 1M | $1.50 | $12.00 |
grok-4.3 |
Exact engine pick | 1M | $1.50 | $3.00 |
gemini-3.5-flash |
Exact engine pick | 1M | $1.80 | $10.80 |
gemini-3.6-flash |
Fast tier default | 1M | $1.80 | $9.00 |
claude-sonnet-5 |
Exact engine pick | 1M | $2.40 | $12.00 |
gpt-4.1 |
Exact engine pick | 1M | $2.40 | $9.60 |
grok-4.20-0309-non-reasoning |
Exact engine pick | 2M | $2.40 | $7.20 |
grok-4.20-0309-reasoning |
Exact engine pick | 2M | $2.40 | $7.20 |
gpt-5.4 |
Exact engine pick | 1M | $3.00 | $18.00 |
gpt-5.6-terra |
Smart tier default | 1.1M | $3.00 | $18.00 |
kimi-k3 |
Exact engine pick | 1M | $3.60 | $18.00 |
claude-opus-4.6 |
Exact engine pick | 1M | $6.00 | $30.00 |
claude-opus-4.7 |
Exact engine pick | 1M | $6.00 | $30.00 |
claude-opus-4.8 |
Exact engine pick | 1M | $6.00 | $30.00 |
claude-opus-5 |
Max tier default | 1M | $6.00 | $30.00 |
gpt-5.5 |
Exact engine pick | 1M | $6.00 | $36.00 |
gpt-5.6-sol |
Exact engine pick | 1.1M | $6.00 | $36.00 |
claude-fable-5 |
Exact engine pick | 1M | $12.00 | $60.00 |
Every price above is paired with how you reach that model: a named tier runs its model by default, and everything else is reached from Pick an exact engine in the composer’s engine menu.
Roughly how much that is
Section titled “Roughly how much that is”Same models, with the window turned into something you can picture. A token is about ¾ of a word and a page about 500 words, so treat these as scale, not arithmetic:
| Model | Context window | Roughly |
|---|---|---|
gemini-2.5-flash-lite |
1M tokens | roughly 1.5k pages |
gpt-4.1-nano |
1M tokens | roughly 1.5k pages |
gemini-3.1-flash-lite |
1M tokens | roughly 1.5k pages |
gemini-2.5-flash |
1M tokens | roughly 1.5k pages |
gemini-3.5-flash-lite |
1M tokens | roughly 1.6k pages |
m3 |
1M tokens | roughly 1.5k pages |
gpt-4.1-mini |
1M tokens | roughly 1.6k pages |
gemini-3-flash |
1M tokens | roughly 1.5k pages |
gpt-5.6-luna |
1.1M tokens | roughly 1.6k pages |
gemini-2.5-pro |
1M tokens | roughly 1.5k pages |
grok-4.3 |
1M tokens | roughly 1.5k pages |
gemini-3.5-flash |
1M tokens | roughly 1.6k pages |
gemini-3.6-flash |
1M tokens | roughly 1.6k pages |
claude-sonnet-5 |
1M tokens | roughly 1.5k pages |
gpt-4.1 |
1M tokens | roughly 1.6k pages |
grok-4.20-0309-non-reasoning |
2M tokens | roughly 3k pages |
grok-4.20-0309-reasoning |
2M tokens | roughly 3k pages |
gpt-5.4 |
1M tokens | roughly 1.5k pages |
gpt-5.6-terra |
1.1M tokens | roughly 1.6k pages |
kimi-k3 |
1M tokens | roughly 1.6k pages |
claude-opus-4.6 |
1M tokens | roughly 1.5k pages |
claude-opus-4.7 |
1M tokens | roughly 1.5k pages |
claude-opus-4.8 |
1M tokens | roughly 1.5k pages |
claude-opus-5 |
1M tokens | roughly 1.5k pages |
gpt-5.5 |
1M tokens | roughly 1.5k pages |
gpt-5.6-sol |
1.1M tokens | roughly 1.6k pages |
claude-fable-5 |
1M tokens | roughly 1.5k pages |
How this list is decided
Section titled “How this list is decided”Context windows come from our own model catalog, which records the window the maker publishes. This page includes every model at 1M tokens or above. Two models on the full list record no window at all; they are not here, because we will not guess one.
A big window is a ceiling, not a promise. It says the request will fit — not that the model will use every part of it well, and not that it will be cheap: you pay per token, so filling a million-token window is a million tokens of input on your receipt. The page-count conversions are rough (a token is about ¾ of a word, a page about 500 words), and they are there for scale, not for planning a bill.
- Model prices — all 60 models, all prices, one table
- The Fast tier, The Smart tier, The Max tier — what you get without choosing a model at all
- Start a run at elsedoes.com/home
Prices last changed 2026-07-29 (source ced6a4f).