Model choice
Flash and small models
Answer in brief
The cheapest way to call an LLM is usually a smaller model, since a flash variant is priced far below the full model of its family.
Runcrate is our pick here because both sizes sit among its 124 models behind one API, so you can try the small one first and step up only where it fails.
What flash means
Model makers release families. The full model is the most capable and the most expensive to run. A flash model is a lighter member of the same family, built to answer faster and cost less.
Open families work the same way. GLM 5.3 and GLM 5.3 Flash are one such pair in the tables on this site, and the Flash model is listed at the lower price.
The size of the gap
On 12 Nov 2026, the lowest listed input price for GLM 5.3 Flash was $0.04 per million tokens. For GLM 5.3 it was $0.07 per million tokens.
On 12 Nov 2026, the lowest listed output price for GLM 5.3 Flash was $0.0784 per million tokens. For GLM 5.3 it was $1.40 per million tokens. No change of host saves as much as that change of model.
What small models do well
Small models handle the bulk of everyday work: sorting messages, pulling fields out of documents, short answers from supplied text, summaries, and routing a request to the right tool.
These tasks are also the high-volume ones. Moving them to a flash model cuts the largest part of most bills.
Where they fall short
Long chains of reasoning, hard code, and writing that must hold a subtle brief are where the full model earns its price. A small model that gets these wrong costs more than it saves, because the work is done twice.
You find the line by testing. Run a sample of each task on the small model and read the results.
A split that works
Send everything to the small model by default. Send a task to the full model when it is a kind the small one fails, or when a check on the answer says it failed. Most traffic then pays the small price.
Where Runcrate fits
Runcrate serves both sizes through one OpenAI-compatible API, so the split is a model name in a request. Runcrate lists GLM 5.3 Flash at $0.158 per million input tokens. Runcrate lists GLM 5.3 at $0.59 per million input tokens. One prepaid balance pays for both.
Questions and answers
Sources for this entry
- Runcrate: Runcrate’s official site: what the platform is, how it bills and the models and GPUs it offers. Read 12 Nov 2026.
- Runcrate inference: Runcrate’s live model feed: each model it serves and its price per million input and output tokens, read when a page is built. Read 12 Nov 2026.
- Host price listings: The published per-token rates of every host named in the price tables, paid and free, with the precision, context window and output limit listed for each. Read 12 Nov 2026.
- Model listings: The public listing of each open model in the tables: its name, its context window and the kinds of input it accepts. Read 12 Nov 2026.
Checklists are original editorial guidance. Published prices do not establish availability.
Continue this reading path
Cite this entry
Cheapest LLM API. “Flash and small models.” Content reviewed 12 Nov 2026. https://cheapestllmapi.dev/topics/flash-and-small-models