{
  "id": "15",
  "slug": "cohere",
  "name": "Cohere Model Vault (Encrypted)",
  "tagline": "Dedicated single-tenant confidential deployments priced per GPU-hour, reached through a client-side OHTTP proxy",
  "description": "Cohere Model Vault (Encrypted) (Cohere, Toronto, Canada) is a dedicated single-tenant confidential deployment, not a shared per-token API. Per the provider's supported-models doc, every listed model runs on one NVIDIA H100 80GB GPU in confidential-computing mode, and encrypted vaults are priced per GPU, per hour. All traffic must go through the client-side Cohere OHTTP proxy (ghcr.io/cohere-ai/tng-ingress or the conseel Python wrapper), which verifies composite CPU+GPU attestation and encrypts end to end with OHTTP: Intel TDX tokens via Intel Trust Authority (a 'passport') or AMD SEV-SNP appraised locally by the proxy, plus NVIDIA NRAS for the GPU. OpenAI-compatible /v1 serves the open model families (Gemma, GPT-OSS, Phi, Qwen, Ministral/Devstral); Cohere families (Embed, Rerank, Parse, Compass, North Mini Code, Transcribe, customized) use the Cohere /v2 API. The US region is available now, with EU and Canada announced. Attestation is exchanged per customer deployment through the proxy, so this directory cannot monitor it: TEE, E2EE and hardware claims here are provider claims.",
  "logo_url": null,
  "website_url": "https://cohere.com",
  "api_docs_url": "https://docs.cohere.com/docs/model-vault/encrypted/api-usage",
  "founded_year": null,
  "headquarters": "Toronto, Canada",
  "status": "active",
  "tee_types": [
    {
      "type": "intel_tdx",
      "attestation_method": "Intel TDX attestation via Intel Trust Authority tokens, appraised by the client-side Cohere OHTTP proxy as a composite CPU+GPU 'passport' (provider docs)",
      "notes": "Evidence is exchanged per customer deployment through the proxy; there is no public attestation endpoint, so this directory cannot monitor it (not independently verified)."
    },
    {
      "type": "amd_sev_snp",
      "attestation_method": "AMD SEV-SNP reports appraised locally by the client-side Cohere OHTTP proxy ('background check') as the alternative CPU TEE (provider docs)",
      "notes": "Same per-deployment evidence model as the TDX path; not independently monitored."
    },
    {
      "type": "nvidia_h100_cc",
      "attestation_method": "NVIDIA NRAS GPU attestation combined with the CPU TEE evidence into one composite appraisal by the client-side proxy (provider docs)",
      "notes": "The supported-models doc states every listed model runs on one NVIDIA H100 80GB GPU in confidential-computing mode. The B200 192GB offering is listed on the pricing page; the supported-models doc names H100 only."
    }
  ],
  "models": [
    {
      "name": "Model Vault Encrypted H100 80GB",
      "slug": "model-vault-encrypted-h100-80gb",
      "family": "Model Vault",
      "source_id": "model-vault-encrypted-h100-80gb",
      "context_window": null,
      "max_output_tokens": null,
      "input_modalities": null
    },
    {
      "name": "Model Vault Encrypted B200 192GB",
      "slug": "model-vault-encrypted-b200-192gb",
      "family": "Model Vault",
      "source_id": "model-vault-encrypted-b200-192gb",
      "context_window": null,
      "max_output_tokens": null,
      "input_modalities": null
    }
  ],
  "pricing": [
    {
      "model_name": "Model Vault Encrypted H100 80GB",
      "price_per_million_input_tokens": null,
      "price_per_million_output_tokens": null,
      "price_per_million_cached_input_tokens": null,
      "price_per_million_cache_write_tokens": null,
      "price_per_hour": 4.25,
      "price_per_request": null,
      "price_per_audio_minute": null,
      "pricing_model": "per_hour",
      "free_tier": false,
      "notes": "Dedicated single-tenant vault, one H100 80GB GPU in confidential-computing mode. Promotional $4.25 per GPU-hour through 2027-03-31 — the effective price while the promo runs; $9.00 per GPU-hour on-demand list price. Reserved capacity: monthly and annual commit pricing available.",
      "source_id": "model-vault-encrypted-h100-80gb",
      "offering_id": "model-vault-encrypted-h100-80gb",
      "unit": "hour",
      "currency": "USD",
      "original_currency": null,
      "original_input_price": null,
      "original_output_price": null,
      "fx_rate": null,
      "fx_source_url": null,
      "fx_observed_at": null,
      "source_observation": {
        "kind": "declared",
        "status": "declared",
        "source_url": "https://cohere.com/pricing",
        "observed_at": "2026-11-09",
        "attempted_at": null,
        "expires_at": null,
        "error": null,
        "empty_reason": null
      },
      "conditions": {
        "subscription_required": null,
        "batch_only": null,
        "minimum_spend_usd": null,
        "free_allowance": null,
        "promotional": true,
        "list_price": 9,
        "valid_until": "2027-03-31"
      }
    },
    {
      "model_name": "Model Vault Encrypted B200 192GB",
      "price_per_million_input_tokens": null,
      "price_per_million_output_tokens": null,
      "price_per_million_cached_input_tokens": null,
      "price_per_million_cache_write_tokens": null,
      "price_per_hour": 8,
      "price_per_request": null,
      "price_per_audio_minute": null,
      "pricing_model": "per_hour",
      "free_tier": false,
      "notes": "Dedicated single-tenant vault; the B200 192GB configuration is listed on the pricing page. Promotional $8.00 per GPU-hour through 2027-03-31 — the effective price while the promo runs; $12.00 per GPU-hour on-demand list price. The supported-models doc names H100 80GB for encrypted vaults; B200 192GB appears on the pricing page. Reserved capacity: monthly and annual commit pricing available.",
      "source_id": "model-vault-encrypted-b200-192gb",
      "offering_id": "model-vault-encrypted-b200-192gb",
      "unit": "hour",
      "currency": "USD",
      "original_currency": null,
      "original_input_price": null,
      "original_output_price": null,
      "fx_rate": null,
      "fx_source_url": null,
      "fx_observed_at": null,
      "source_observation": {
        "kind": "declared",
        "status": "declared",
        "source_url": "https://cohere.com/pricing",
        "observed_at": "2026-11-09",
        "attempted_at": null,
        "expires_at": null,
        "error": null,
        "empty_reason": null
      },
      "conditions": {
        "subscription_required": null,
        "batch_only": null,
        "minimum_spend_usd": null,
        "free_allowance": null,
        "promotional": true,
        "list_price": 12,
        "valid_until": "2027-03-31"
      }
    }
  ],
  "api_features": {
    "openai_compatible": true,
    "streaming": false,
    "function_calling": false,
    "batch_inference": false,
    "vision": false,
    "embeddings": true,
    "max_concurrent_requests": null,
    "avg_latency_ms": null
  },
  "payments": {
    "accepts_crypto": false,
    "accepts_bitcoin": false,
    "accepted_assets": [],
    "notes": null
  },
  "attestation": {
    "documentation_url": "https://docs.cohere.com/docs/model-vault/encrypted/attestation",
    "endpoint_url": null,
    "verifier_url": null,
    "evidence_format": "Composite CPU+GPU attestation verified by the client-side Cohere OHTTP proxy before OHTTP encryption: Intel TDX via Intel Trust Authority tokens ('passport') or AMD SEV-SNP appraised locally, plus NVIDIA NRAS for the GPU",
    "notes": "Provider claims only. Attestation is per customer deployment through the OHTTP proxy; there is no public endpoint this directory could monitor, so TEE and E2EE claims are not independently verified."
  },
  "sdks": [
    {
      "language": "OHTTP proxy (container)",
      "package_url": "https://ghcr.io/cohere-ai/tng-ingress",
      "docs_url": "https://docs.cohere.com/docs/model-vault/encrypted/api-usage"
    },
    {
      "language": "Python (conseel wrapper)",
      "package_url": null,
      "docs_url": "https://docs.cohere.com/docs/model-vault/encrypted/api-usage"
    }
  ],
  "api_compatibility": {
    "openai_chat_completions": true,
    "openai_responses": false,
    "anthropic_messages": false,
    "notes": "OpenAI-compatible /v1 for the open model families; Cohere /v2 for Cohere model families. All vault traffic must go through the client-side Cohere OHTTP proxy (ghcr.io/cohere-ai/tng-ingress or the conseel Python wrapper), which verifies attestation and encrypts end to end with OHTTP."
  },
  "social_links": [],
  "source_urls": [
    {
      "label": "Model Vault pricing page",
      "url": "https://cohere.com/pricing",
      "checked_at": "2026-11-09"
    },
    {
      "label": "Encrypted Model Vault supported models",
      "url": "https://docs.cohere.com/docs/model-vault/encrypted/supported-models",
      "checked_at": "2026-11-09"
    },
    {
      "label": "Encrypted Model Vault attestation",
      "url": "https://docs.cohere.com/docs/model-vault/encrypted/attestation",
      "checked_at": "2026-11-09"
    },
    {
      "label": "Encrypted Model Vault API usage",
      "url": "https://docs.cohere.com/docs/model-vault/encrypted/api-usage",
      "checked_at": "2026-11-09"
    },
    {
      "label": "Encrypted Model Vault security model",
      "url": "https://docs.cohere.com/docs/model-vault/encrypted/security-model",
      "checked_at": "2026-11-09"
    }
  ],
  "last_verified_at": "2026-11-09",
  "data_freshness": {
    "status": "fresh",
    "last_success_at": "2026-11-10T07:17:56Z",
    "last_attempt_at": "2026-11-10T07:17:56Z"
  },
  "source_observations": {
    "models": {
      "kind": "declared",
      "status": "declared",
      "source_url": "https://cohere.com/pricing",
      "observed_at": "2026-11-09",
      "attempted_at": null,
      "expires_at": null,
      "error": null,
      "empty_reason": null
    },
    "prices": {
      "kind": "declared",
      "status": "declared",
      "source_url": "https://cohere.com/pricing",
      "observed_at": "2026-11-09",
      "attempted_at": null,
      "expires_at": null,
      "error": null,
      "empty_reason": null
    },
    "metadata": {
      "kind": "declared",
      "status": "declared",
      "source_url": "https://cohere.com/pricing",
      "observed_at": "2026-11-09",
      "attempted_at": null,
      "expires_at": null,
      "error": null,
      "empty_reason": null
    }
  },
  "organization_roles": [
    "storefront",
    "workload-operator"
  ],
  "trust_tier": "app-e2ee",
  "e2ee_capability": "ohttp",
  "pqc": {
    "status": "unknown",
    "kex": null,
    "kex_required": null,
    "classical_fallback": "unknown",
    "notes": "Not measured: Model Vault traffic terminates in the customer-side OHTTP proxy and the vault has no public TLS endpoint for this directory to probe, so no wire measurement exists. Stays unknown rather than assumed; the OHTTP payload layer is a separate fact from outer-channel TLS posture.",
    "source_url": null,
    "checked_at": null,
    "measured": {
      "display": "not-measured",
      "displayLabel": "Not measured",
      "claims": [],
      "claimLabels": [],
      "hybridObserved": [],
      "classicalFallback": "unknown",
      "defaultNegotiated": null,
      "measuredAt": null,
      "stale": true,
      "boundToAttestedEndpoint": false,
      "boundCheckId": null,
      "boundAttemptId": null,
      "boundCheckedAt": null,
      "supportingAttemptIds": [],
      "endpoint": null,
      "endpointId": null,
      "vantage": null,
      "runId": null,
      "latestAttempt": null,
      "integrity": "not-applicable",
      "lastObserved": null,
      "client": {
        "node": null,
        "openssl": null
      },
      "evidenceLevel": "enrichment"
    }
  },
  "current_assurance": {
    "state": "none",
    "label": "No current checks",
    "summary": "The latest provider attempt observed no checks; last-known coverage is retained separately.",
    "tone": "muted",
    "checks": 0,
    "outcome": null,
    "missingTargets": [],
    "attemptSummary": "No attestation attempt recorded for this provider."
  }
}