Skip to content
GridRouterhome

Search

Search providers, capabilities and pages

Cohere API

cohere.com

AI UtilitiesListed, not routed yet
Command, Embed, Rerank and Parse model APIs for enterprise.

What it does

Enterprise LLM provider (Command chat, Embed, Rerank, Parse, Aya). Trial keys are free but rate limited and limited to 1,000 calls/month; production keys have higher per-endpoint limits (e.g. Command A 500 req/min, Rerank 1,000 req/min). Model Vault dedicated instances priced hourly/monthly.

Pricing

What Cohere publishes for buying direct, as listed on its own pages.

Buying direct from Cohere

  • Command (legacy)$1.00/1M input tokens
  • Model Vault Embed 4 Small$4.00/hour per instance

Free tier: Trial key: 1,000 API calls a month, non-production. Larger plans through sales. Cohere pricing, checked Sep 29, 2026

Docs

Cohere's own pages, linked as published.

Its OpenAPI document: 42 operations and rate limits
Official specOpenAPI 3.1.0 · 42 operations Spec (YAML)Provider's spec
Operations in the cohere spec
MethodPathSummaryParamsSource
POST/v1/chatChat API (v1)1Spec
POST/v2/chatChat API (v2)0Spec
POST/v2/parseParse API (v2)0Spec
POST/v1/generateGenerate0Spec
POST/v1/embedEmbed API (v1)0Spec
POST/v2/embedEmbed API (v2)0Spec
POST/v2/audio/transcriptionsCreate a transcription0Spec
GET/v2/batchesList batches3Spec
POST/v2/batchesCreate a batch0Spec
GET/v2/batches/{id}Retrieve a batch1Spec
POST/v2/batches/{id}:cancelCancel a batch1Spec
GET/v1/embed-jobsList Embed Jobs0Spec
POST/v1/embed-jobsCreate an Embed Job0Spec
GET/v1/embed-jobs/{id}Fetch an Embed Job1Spec
POST/v1/embed-jobs/{id}/cancelCancel an Embed Job1Spec
POST/v1/rerankRerank API (v1)0Spec
POST/v2/rerankRerank API (v2)0Spec
POST/v1/classifyClassify0Spec
GET/v1/datasetsList Datasets6Spec
POST/v1/datasetsCreate a Dataset8Spec
GET/v1/datasets/usageGet Dataset Usage0Spec
GET/v1/datasets/{id}Get a Dataset1Spec
DELETE/v1/datasets/{id}Delete a Dataset1Spec
POST/v1/summarizeSummarize0Spec
POST/v1/tokenizeTokenize0Spec
POST/v1/detokenizeDetokenize0Spec
GET/v1/connectorsList Connectors2Spec
POST/v1/connectorsCreate a Connector0Spec
GET/v1/connectors/{id}Get a Connector1Spec
PATCH/v1/connectors/{id}Update a Connector1Spec
DELETE/v1/connectors/{id}Delete a Connector1Spec
POST/v1/connectors/{id}/oauth/authorizeAuthorize with oAuth2Spec
GET/v1/models/{model}Get a Model1Spec
GET/v1/modelsList Models4Spec
POST/v1/check-api-keyCheck API key0Spec
GET/v1/finetuning/finetuned-modelsLists fine-tuned models.3Spec
POST/v1/finetuning/finetuned-modelsTrains and deploys a fine-tuned model.0Spec
GET/v1/finetuning/finetuned-models/{id}Returns a fine-tuned model by ID.1Spec
PATCH/v1/finetuning/finetuned-models/{id}Updates a fine-tuned model.1Spec
DELETE/v1/finetuning/finetuned-models/{id}Deletes a fine-tuned model.1Spec

Showing 40 of 42 operations; the full list is in the spec.

Rate limits

Published rate limits for cohere
LimitScopePlanEndpointBurst / concurrency
20 / minutekeyTrialChat (Command A)— / —
500 / minutekeyProductionChat (Command A)— / —
10 / minutekeyTrialRerank— / —
1,000 / minutekeyProductionRerank— / —
2,000 / minutekeyAllEmbed— / —
1,000 / monthkeyTrialAll— / —

GridRouter never paces above these and backs off on 429, honoring Retry-After. Source, checked Sep 29, 2026

Sources for this page (3)