High-value primary-source verification.

providerApi: True, baseUrl: https://api.openai.com/v1

GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna

reasoning, code-generation, image-understanding +4

OpenAI API, ChatGPT business/consumer product surfaces

licenseClass: proprietary_service, commercialUse: service-terms-dependent

gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna

businessDataTrainingDefault: False, encryptionAtRest: AES-256

providerApi: True, baseUrl: https://api.anthropic.com

Claude Fable 5.1, Claude Opus 5, Claude Sonnet 5 +1

reasoning, agentic-planning, code-generation +4

Claude API, Claude Platform on AWS, Amazon Bedrock +2

licenseClass: proprietary_service, commercialUse: service-terms-dependent

claude-fable-5-1, claude-opus-5, claude-sonnet-5 +1

trainingOnRetainedApiContentWithoutPermission: False, zeroDataRetention: available for eligible API arrangements/features

providerApi: True, baseUrl: https://generativelanguage.googleapis.com/v1beta

Gemini 3.7 Flash

reasoning, code-generation, agentic-planning +3

Gemini Developer API, Google AI Studio, Google Cloud / Vertex AI enterprise path

licenseClass: proprietary_service, commercialUse: service-terms-dependent

gemini-3.7-flash, gemini-3.7-flash

billingEnabledApiLogsTrainingDefault: False, apiLoggingRetentionMaxDays: 55

providerApi: True, baseUrl: https://api.x.ai

Grok 4.6

reasoning, code-generation, agentic-planning +5

SpaceXAI/xAI API, Grok Build, Cursor +1

licenseClass: proprietary_service, commercialUse: service-terms-dependent

grok-4.6, grok-4.6

apiTrainingWithoutExplicitPermission: False, defaultRetentionDays: 30

providerApi: True, baseUrl: https://api.deepseek.com

DeepSeek V4 Pro, DeepSeek V4 Flash

reasoning, agentic-planning, function-calling +2

DeepSeek API, DeepSeek app/web

Needs verification

deepseek-v4-flash, deepseek-v4-flash, deepseek-v4-pro +1

Needs verification

providerApi: True, baseUrl: https://api.mistral.ai/v1

Mistral Small 4

reasoning, code-generation, function-calling +3

Mistral hosted API, Vibe product, self-host/open-weight option for selected models

model: Mistral Small 4, licenseClass: open_weights_permissive

Mistral Small 4, Mistral Large 3, Mistral Medium 3.5

vibeFreeTraining: may be used; opt-out available, vibeProTeamEnterpriseTrainingDefault: False

providerApi: True, authentication: ['Authorization: Bearer <token>']

Command A+, Command A

reasoning, multilingual, image-understanding +4

Cohere API, Cohere Model Vault

licenseClass: proprietary_or_commercial_deployment, commercialUse: supported through production offerings

Command A, Command A+

privateDeploymentDataAccessibleToCohere: False, enterpriseTrainingControls: True

providerApi: True, interface: Amazon Bedrock runtime APIs

Nova 2 Lite, Nova 2 Sonic, Nova Multimodal Embeddings

reasoning, long-context, image-understanding +8

Amazon Bedrock managed inference, SageMaker customization paths for supported Nova models

licenseClass: proprietary_service, commercialUse: AWS service terms/EULA dependent

pricingModel: model-specific Amazon Bedrock usage pricing, exactRates: None

customerInputsOutputsUsedToTrainBaseModels: False, customerInputsOutputsSharedWithThirdPartyModelProviders: False

Needs verification

Cursor Grok 4.6, Cursor Grok 4.5, Composer 2.5 +6

code-generation, agentic-planning, repository-analysis +1

Cursor editor, Cloud Agents, provider/trusted-partner/Cursor-hosted model routing

licenseClass: proprietary_service, commercialUse: paid professional/team/enterprise plans available

{'plan': 'Pro', 'unit': 'USD/month', 'price': 20}, {'plan': 'Pro Plus', 'unit': 'USD/month', 'price': 60}, {'plan': 'Ultra', 'unit': 'USD/month', 'price': 200} +1

privacyModeTrainingByCursor: False, privacyModeProviderTraining: False

Needs verification

GPT-5 mini, GPT-5.3-Codex, GPT-5.4 +12

code-generation, code-completion, agentic-planning +2

GitHub.com, IDE integrations, Copilot CLI +1

licenseClass: proprietary_service, businessAndEnterprisePlans: True

{'plan': 'Pro', 'unit': 'USD/month', 'price': 10}, {'plan': 'Pro+', 'unit': 'USD/month', 'price': 39}, {'plan': 'Max', 'unit': 'USD/month', 'price': 100} +3

businessEnterpriseCustomerDataUsedForTraining: False, individualInteractionDataMayBeUsedForTraining: True

providerApi: True, authentication: ['organization-scoped API key']

{'id': 'gen4.5', 'type': 'video', 'flagship': True}, {'id': 'gen4_turbo', 'type': 'video'}, {'id': 'aleph2', 'type': 'video_editing'} +2

video-generation, video-editing, image-generation +3

Runway Dev API, Runway cloud creative product

Needs verification

{'unit': 'USD/credit', 'price': 0.01}, gen4.5, gen4_turbo +1

soc2Type2: True, uploadedAssetsPrivateByDefault: True

providerApi: True, baseUrl: https://api.elevenlabs.io/v1

Eleven v3, Eleven v3 Conversational, Eleven Flash v2.5 +3

speech-synthesis, speech-recognition, speaker-diarization +3

ElevenLabs API, web products, isolated regional environments for eligible customers

Needs verification

{'product': 'TTS v3', 'unit': 'USD/1K characters', 'price': 0.1}, {'product': 'TTS Flash/Turbo', 'unit': 'USD/1K characters', 'price': 0.05}, {'product': 'Scribe STT', 'unit': 'USD/audio-hour', 'price': 0.22} +2

zeroRetentionMode: select enterprise API use; most eligible request/response content deleted after request, thirdPartyLlmTrainingOnCustomerContent: False

providerApi: True, endpoint: https://api.pinecone.io/assistant/assistants

Needs verification

retrieval, question-answering, document-search +1

Pinecone managed cloud Assistant

licenseClass: proprietary_managed_service, commercialUse: paid Standard/Enterprise service available

{'plan': 'Standard/Enterprise', 'unit': 'USD/1M assistant input tokens', 'price': 8}, {'plan': 'Standard/Enterprise', 'unit': 'USD/1M assistant output tokens', 'price': 15}, {'plan': 'Standard/Enterprise', 'unit': 'USD/GB/month assistant storage', 'price': 3}

assistantRegions: ['us', 'eu'], encryptionAtRest: AES-256

providerApi: True, freeEndpoint: https://api-free.deepl.com

Needs verification

translation, document-analysis, editing +2

DeepL API, DeepL web/desktop products

licenseClass: proprietary_service, commercialUse: commercial API plans available

Developer: 1,000,000 characters total before upgrade, GrowthMonthlyIncluded: {'characters': 1000000, 'speechToTextHours': 10}

paidProductContentTrainingWithoutConsent: False, proApiTextsPermanentlyStored: False