The home of open models, datasets and Spaces.
Cohere
Enterprise models for generation, embeddings and rerank.
SharkScore
#9 in Model APIs & inference
Why it scores 72.1
Cohere has a SharkScore of 72.1/100, ranked #9 of 262 in Model APIs & inference (updated 2026-11-18).
Each part compares Cohere with other listings in Model APIs & inference. Evidence covers 87% of the score's weight.
- Adoption
- 57 · 30%
- User reviews
- no data · 25%
- Maintenance
- 100 · 20%
- Trust & transparency
- 80 · 15%
- Momentum
- no data · 10%
Paying never changes this score. How ranking works · method v2026-11-18.2
RankShark review of Cohere
Automated evaluation by RankShark · 18 Nov 2026 · method v1.0 · How we review
We were not paid for this review.
As of 2026-11-18, Cohere offers enterprise AI models and platforms, with free but rate-limited trial API access and paid production API use billed pay-as-you-go. [E1, E2]
Cohere offers generative language, speech recognition, translation, document parsing, embedding, and reranking models, plus enterprise products such as North, Compass, and Model Vault. It provides API access and private deployment options. [E1, E2, E3, E6]
Best for
- Enterprise teams building AI applications with model APIs, search, or retrieval features. [E1, E2, E3] · home page, pricing page, docs
- Organizations seeking private or dedicated deployment options. [E2, E6] · pricing page, security page
- Developers who want to test API use cases with a free trial key. [E2, E3] · pricing page, docs
Not for
- Users who need a stated fixed monthly price for API access; production API use is billed pay-as-you-go, and no fixed monthly API price is stated. [E2] · pricing page
- Users who need to use a trial API key for production or commercial purposes; the pricing page says trial keys are not permitted for those uses. [E2] · pricing page
Key features
- Model catalog. The vendor lists Command generative language models, Transcribe speech recognition, North Small Translate machine translation, Parse document parsing, Embed semantic representation, and Rerank retrieval optimization. [E1] · home page
- Enterprise platforms. The vendor describes North as an enterprise AI platform and Compass as an intelligent search and discovery system. [E2] · pricing page
- Model hosting. The vendor says Model Vault is a managed platform for running Cohere and open-weight models, with pricing based on GPUs and replica count. [E2] · pricing page
- Deployment choices. The vendor lists VPC, on-premises, and Cohere-managed Model Vault deployment options. [E6] · security page
- Developer resources. The vendor provides API documentation, release notes, cookbooks, a no-code Playground, and Discord community access. [E3] · docs
Pros
- The product lineup includes language, speech, translation, parsing, embedding, and reranking models. [E1] · home page
- The vendor lists private deployment options including VPC, on-premises, and Model Vault. [E6] · security page
- A free, rate-limited API trial is available without a stated payment requirement. [E2, E4] · pricing page, privacy page
Cons
- Trial API keys are rate limited and cannot be used for production or commercial purposes. [E2] · pricing page
- The pricing page gives custom enterprise pricing for North and Compass, and production API access is pay-as-you-go rather than a stated fixed monthly price. [E2] · pricing page
- The privacy policy says trial and research inputs and outputs may be used for research and development. [E4] · privacy page
Pricing
| Plan | Price | Limits |
|---|---|---|
| North | Custom enterprise pricing. | The pricing page does not state a fixed price or usage limit. |
| Compass | Custom enterprise pricing. | The pricing page does not state a fixed price or usage limit. |
| Model Vault — L4 24GB* | $3.00/GPU/hour on-demand; $1.00/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Model Vault — A100 40GB | $3.25/GPU/hour on-demand; $2.25/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Model Vault — A100 80GB | $3.50/GPU/hour on-demand; $2.25/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Model Vault — MI300X 192GB | $6.00/GPU/hour on-demand; $4.00/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Model Vault — H100 80GB | $6.25/GPU/hour on-demand; $4.25/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Model Vault — H200 141GB | $7.25/GPU/hour on-demand; $5.50/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Model Vault — H100 80GB (encrypted) | $9.00/GPU/hour on-demand; $4.25/GPU/hour promotional through March 31 ‘27. | The table marks this option as Model Vault Encrypted (with Confidential Compute); pricing is per GPU per hour. |
| Model Vault — B200 192GB | $9.75/GPU/hour on-demand; $8.00/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Model Vault — B200 192GB (encrypted) | $12.00/GPU/hour on-demand; $8.00/GPU/hour promotional through March 31 ‘27. | The table marks this option as Model Vault Encrypted (with Confidential Compute); pricing is per GPU per hour. |
| Model Vault — GB200 192GB | $12.50/GPU/hour on-demand; $10.00/GPU/hour promotional through March 31 ‘27. | Pricing is per GPU per hour; the page says pricing does not indicate GPU availability. |
| Trial API key | Free. | Trial keys are rate limited and are not permitted for production or commercial purposes. |
| Production API key | Pay-as-you-go. | The pricing page does not state a fixed monthly price for production API access. |
Checked 18 Nov 2026 · source
- Free plan
- Trial API calls are free, but trial keys are rate limited and not permitted for production or commercial purposes. · pricing page
- Free trial
- An account receives a trial API key; calls are free, rate limited, and not permitted for production or commercial purposes. · pricing page
- Trains on customer data
- The privacy policy says trial users' and researchers' inputs and outputs may be used for research and development. For enterprise users, the policy says they can control how Cohere trains on their data. The policy also says Cohere does not access or use data from private and third-party deployments. · privacy page
- Data retention
- The security page says data handling and retention settings are customizable, but does not specify retention periods. · security page
- Data regions
- Not stated (checked 18 Nov 2026)
- Data processing agreement
- The privacy policy says enterprise users can request a DPA. · privacy page
- AI models used
- The vendor lists Command, Transcribe, North Small Translate, Parse, Embed, Rerank, and Aya models. [E1] · home page
- API
- The vendor offers API access, an API reference, and trial and production API keys. [E2, E3] · pricing page, docs
- SDKs
- Not stated (checked 18 Nov 2026)
- Webhooks
- Not stated (checked 18 Nov 2026)
- Platforms
- The vendor lists VPC, on-premises, Model Vault, and third-party cloud AI/ML deployments, including OCI, Azure, AWS, and Google Cloud. [E6] · security page
- Security claims
- Vendor states the Cohere API platform is SOC 2 Type II compliant. · security pageVendor says it conducts annual third-party audits and penetration tests. · security pageVendor says its API platform uses advanced threat detection systems to monitor for suspicious activity. · security pageVendor says private deployment customers retain control of their infrastructure and data, which Cohere cannot access. · security page
What we measured
- GitHub stars · now
- 409
- Latest release · now
- 7.3.2 on 2026-11-04
- Package downloads · 30 days
- 14,017,551
- Tranco traffic rank · list of the day
- #12,623
How we tested · sources
- Hourly uptime and response-time checks of the product's URL.
- Repository, release and download data from GitHub, npm, PyPI and Hugging Face where they exist.
- We did not sign up, pay, or use the product as a customer.
- home page: https://cohere.com/ · fetched 18 Nov 2026
- pricing page: https://cohere.com/pricing · fetched 18 Nov 2026
- docs: https://cohere.com/developers · fetched 18 Nov 2026
- privacy page: https://cohere.com/privacy · fetched 18 Nov 2026
- terms: https://cohere.com/legal · fetched 18 Nov 2026
- security page: https://cohere.com/security · fetched 18 Nov 2026
Something wrong? Report an error with a public link that shows it; corrections re-run the review.
Top alternatives to Cohere
See all- 78.9ToolFreemium · from $9/mo
- 77.5
API for building with Claude models.
ToolPaid - 76.1
Proxy and SDK to call 100+ LLM APIs in the OpenAI format.
ToolOpen source - 75.7
Google Cloud's platform to build with Gemini and other models.
ToolPaid