OpenAI
LLM18 models / endpointsOpenAI provides hosted language-model APIs billed primarily by token usage. This page connects the verified catalog entries to practical monthly cost planning.
Every service the calculator can price, how it bills, and where coverage is not complete enough to calculate.
Snapshot 2026-10-08-7b05c4ce
OpenAI provides hosted language-model APIs billed primarily by token usage. This page connects the verified catalog entries to practical monthly cost planning.
Anthropic offers hosted Claude models for language workloads. The calculator prices the exact catalogued model from input and output tokens rather than a flat request average.
Google Gemini Developer API supplies hosted language models with token-based billing. The calculator keeps model rates and limits in a versioned pricing snapshot.
Tavily provides search APIs for agent and research workflows. The calculator converts monthly searches into the verified credit-based billing rule for the catalogued profile.
Exa supplies search APIs for retrieval and research applications. The active catalog entry prices a specific request profile without assuming optional content operations.
Firecrawl turns web pages into content for research and retrieval workflows. The calculator models its catalogued monthly package and overage behavior.
ScraperAPI provides managed web requests for scraping workflows. The catalog models a standard request profile and its monthly hard-cap plan.
fal.ai hosts image and video generation endpoints with model-specific output pricing. Each compatible model can be opened as a locked calculator scenario.
Replicate hosts independently priced image and video models. The calculator treats each model and output profile as its own offer instead of using one platform-wide rate.
Apify is relevant to web extraction, but its workload-dependent resource model is not represented as a universal per-page price in this MVP.
OpenRouter aggregates access to many language models and hosts. This MVP does not reduce its changing model-level catalog to one representative token price.
DeepSeek supplies hosted language models with token-based billing. These estimates use peak-hour, uncached rates; off-peak and cache-hit requests can cost less.
TypeSafe AI provides an API for structured decisions. Its Jev model evaluates text against questions and returns answers your software can act on.