Buying snapshot
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Decision point | What Cohere states |
|---|---|
Deployment | Private deployment of all products on-premises or in isolated VPCs, or dedicated Model Vault instances managed by Cohere. |
Cloud access | The Python SDK reaches Cohere models on the Cohere platform, AWS (Bedrock, SageMaker), Azure, GCP and Oracle OCI. |
Customization | Bespoke model customization for enterprise customers, priced to their needs. |
Compliance | SOC 2, GDPR, CCPA, ISO 27001 and Cyber Essentials. |
Our take
Choose Cohere when your organization needs language, retrieval and agent capabilities that it can run privately or on dedicated capacity, and is prepared to buy the North or Compass platforms through sales. Teams that only need metered API calls should note that Cohere labels its published token rates as rates for existing customers.
Evaluate Cohere
Comparisons
Where it stands out
Dedicated, isolated inference. Model Vault runs Cohere models with no shared resources or multi-tenancy overhead, end-to-end encryption and guaranteed performance, and it integrates with North.
A full retrieval stack. Embed, Rerank and Parse models cover embeddings, result reranking and document parsing, and Compass adds pre-built data connectors, intelligent search, document parsing and a managed index.
Product screenshot
Trade-offs to weigh
Free trial keys cannot go to production. Trial API calls are free but rate-limited and not permitted for production or commercial use. A production key needs an account Owner to complete the Go to Production application, and organizations do not receive production access automatically.
Not every API is on every cloud. On OCI on-demand inference, Embed and Chat are available, but Generate and Rerank models must be fine-tuned and deployed to a dedicated endpoint first.
Pricing and plan conditions
North and Compass have custom enterprise pricing. Production API calls are pay-as-you-go, with a bill issued at the end of each calendar month or when the outstanding balance reaches $250. Model Vault is priced per instance by model and performance tier, billed hourly or through monthly or annual commitments, with Fixed or Flex plans.
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Model | Performance tier | Hourly rate | Monthly rate |
|---|---|---|---|
Embed 5 Fast | Small | $3.00 | $2,000 |
Embed 5 Fast | Medium | $5.00 | $3,250 |
Embed 5 Pro | Small | $3.00 | $2,000 |
Embed 5 Pro | Medium | $5.00 | $3,250 |
Rerank 4 Fast | Medium | $5.00 | $3,250 |
Rerank 4 Pro | Medium | $5.00 | $3,250 |
Rerank 4 Pro | Large | $10.00 | $6,500 |
Rerank 3.5 | Medium | $5.00 | $3,250 |
Parse 5 | Medium | $4.00 | $2,500 |
Parse 5 | XL | $7.00 | $4,300 |
Token rates published for existing customers (per 1 million input / output tokens): Command $1.00 / $2.00; Command-light $0.30 / $0.60; Command R 03-2024 $0.50 / $1.50; Command R+ 04-2024 $3.00 / $15.00; Command R+ 08-2024 $2.50 / $10.00; Aya Expanse 8B and 32B $0.50 / $1.50.
Rerank search unit: one query with up to 100 documents; any document over 500 tokens, including the query, is split into chunks that each count as a document.
Public signals
About these signals
Verified factual signals from public sources. They indicate observable activity or interest, not total adoption, product quality, or cost.
6 GitHub commits 90d406 GitHub stars0 vulnerabilities across 2 packages
See all signals from 8 sources
PyPI · cohere@7.3.0 · npm · cohere-ai@8.3.0
Alternatives to Cohere
The reviewed substitutes for Cohere among the foundation model providers, and what would make each one the better answer.
Direct alternatives
Reviewed substitutes: products bought for the same job, where a team picks one.
- Gemini API
- Two foundation-model APIs an application calls for text generation and related model tasks, each sold as per-token API access to the vendor's models (Google's Gemini Developer API pricing page lists paid rates per 1M input and output tokens). A team routes its traffic to one provider or splits it deliberately.Applies to: Choosing which foundation-model API an application will call.
- OpenAI
- OpenAI and Cohere offer language models through an API for text generation, embeddings and retrieval-augmented applications. OpenAI's GPT models span text, code, vision and audio for a broad market, while Cohere aims at enterprises, with models for search, embeddings and classification and data privacy controls built in.Applies to: Choosing between two products of the same kind for one job.
Other approaches
A different approach to the same problem. Each substitutes only for the workload named beside it.
- Anthropic
- Anthropic is chosen instead of Cohere for AI application workloads where reliability, interpretability, and steerability are the primary model-selection requirements. **Anthropic** is an AI research and product company focused on reliable, interpretable, and steerable AI systems.
- Mistral AI
- Cohere is chosen instead of Mistral AI for enterprise retrieval-augmented generation and search-ranking workloads. **Cohere** is an enterprise AI platform for production language-model applications, with text generation, embeddings, retrieval, classification, and reranking capabilities.
- Together AI
- Cohere and Together AI both give an application hosted language models through an API, so either can supply the model layer of an AI stack. Cohere sells its own enterprise models for generation, embeddings and retrieval with data privacy controls, while Together AI runs and fine-tunes open-source models on serverless endpoints or dedicated GPU clusters. The choice turns on whether the stack is built on a vendor's own models or on open models.Applies to: When an application calls hosted language models through an API instead of running models in-house.