# Kong AI Gateway vs LiteLLM

**Leader by Vioscale score:** LiteLLM

| Attribute | Kong AI Gateway | LiteLLM |
|---|---|---|
| **Vioscale score** | 55.8 (36% (low)) | 81 (36% (low)) |
| deployment.options | `{"cloud":true,"hybrid":true,"self_hosted":true}` | `{"cloud":true,"hybrid":true,"on_prem":true,"air_gapped":true,"self_hosted":true}` |
| description.long | Kong AI Gateway is a control platform that routes and governs all AI-related traffic—including language model requests, agent-to-agent communication, and model context protocol interactions—with unified security, monitoring, and compliance enforcement across an organization's AI infrastructure. | An open-source AI gateway that provides a centralized control point for managing multiple large language models from different providers. It offers virtual key management, spend tracking, budget controls, and intelligent request routing without charging per token. |
| features.capabilities | `{"caching":"semantic","open_source":true,"rate_limiting":true,"automatic_failover":true,"budget_spend_controls":true,"openai_compatible_api":true,"virtual_key_management":true,"request_cost_observability":true,"mit_or_apache_permissive_oss_license":true,"pii_redaction_and_data_masking_in_flight":true,"custom_rate_limiting_by_user_or_tenant_id":true,"auto_fallback_routing_on_rate_limit_or_downtime":true,"organization_wide_cost_tracking_and_chargebacks":true,"semantic_caching_for_repeat_queries_and_cost_savings":true}` | `{"open_source":true,"pricing_model":"open-source-self-host","rate_limiting":true,"deployment_model":"self-hosted","routing_strategy":"cost-based","architecture_model":"open_source_proxy","automatic_failover":true,"budget_spend_controls":true,"guardrails_moderation":"built-in","openai_compatible_api":true,"virtual_key_management":true,"request_cost_observability":true,"bring_your_own_keys_byok_routing":true,"mit_or_apache_permissive_oss_license":true,"custom_rate_limiting_by_user_or_tenant_id":true,"universal_api_key_for_100_plus_llm_providers":true,"auto_fallback_routing_on_rate_limit_or_downtime":true,"organization_wide_cost_tracking_and_chargebacks":true}` |
| integrations.count | 16 | 140 |
| integrations.list | `[{"name":"OpenAI"},{"name":"Anthropic"},{"name":"AWS Bedrock"},{"name":"GCP Gemini"},{"name":"Azure AI"},{"name":"Databricks"},{"name":"Mistral"},{"name":"Huggingface"},{"name":"Llama"},{"name":"AWS Bedrock Guardrails"},{"name":"Azure Content Safety"},{"name":"Google Cloud Model Armor"},{"name":"Prometheus"},{"name":"Datadog"},{"name":"Grafana"},{"name":"Git"},{"name":"CertManager"},{"name":"Google Cloud Gemini"},{"name":"Hugging Face"},{"name":"Docusign"},{"name":"Kafka"},{"name":"AWS Kinesis"},{"name":"Google Pub/Sub"},{"name":"RabbitMQ"}]` | `[{"name":"OpenAI"},{"name":"Anthropic"},{"name":"Azure"},{"name":"AWS Bedrock"},{"name":"Cloudflare"},{"name":"Google VertexAI"},{"name":"Meta Llama"},{"name":"Mistral AI"},{"name":"NLP Cloud"},{"name":"OVHCloud"},{"name":"vLLM"},{"name":"Nvidia NIM"},{"name":"Slack"},{"name":"Discord"},{"name":"Microsoft Teams"},{"name":"PagerDuty"},{"name":"Prometheus"},{"name":"AWS Secrets Manager"},{"name":"Google Cloud Secret Manager"},{"name":"MCP servers"},{"name":"Gemini"},{"name":"Bedrock"},{"name":"VertexAI"},{"name":"Langfuse"},{"name":"Langsmith"},{"name":"Datadog"},{"name":"Email"},{"name":"Webhook"},{"name":"S3"},{"name":"Google Cloud Storage"},{"name":"Azure Data Lake"},{"name":"GCP Secret Manager"}]` |
| market.availability | `{"primaryMarkets":[],"availabilityScope":"global","availableCountries":[],"notAvailableCountries":[]}` | `{"primaryMarkets":[],"availabilityScope":"global","availableCountries":[],"notAvailableCountries":[]}` |
| platform.support | `{"cli":true,"web":true,"linux":true}` | `{"cli":true,"web":true}` |
| pricing | `{"type":"hybrid","currency":"USD","freeTier":true,"sourceUrl":"https://konghq.com/products/kong-ai-gateway","retrievedAt":"2026-08-21T13:45:38.191Z"}` | `{"type":"hybrid","plans":[{"free":true,"name":"Open Source","summary":"Free forever, self-hosted","features":["100+ LLM providers","Virtual keys, users & teams","Spend tracking","Prometheus metrics","Logging and alerting","Admin UI","Guardrails","Load balancing"],"components":[{"kind":"fixed","amount":0,"period":"month","currency":"USD"}],"description":"MIT-licensed, self-hosted gateway","contactSales":false,"includedLimits":{"providers":"100+","deployment":"self-hosted"}},{"free":false,"name":"Enterprise","summary":"Priced annually based on gateway request capacity and deployment architecture","features":["Everything in Open Source, plus:","SSO + SCIM provisioning","OIDC / JWT authentication","Audit logs on every request","Secret manager integration","Key rotation","Organization & team administration","Multi-region control plane","24/7 support & SLAs","Dedicated support & onboarding","Volume discount tiers","Air-gapped deployment available"],"commitment":"annual","description":"Production deployment with governance, security, and professional support","contactSales":true,"includedLimits":{"deployment":"self-hosted or air-gapped"}}],"summary":"Free open-source tier. Enterprise pricing based on annual request volume, not per-token.","currency":"USD","freeTier":true,"sourceUrl":"https://www.litellm.ai/pricing","retrievedAt":"2026-08-19T13:43:01.911Z","freeTrialDays":30,"billingPeriods":["year"]}` |
| pricing.free_tier | yes | yes |
| pricing.model | freemium | freemium |
| pricing.price_level | low | low |
| pricing.transparent | no | yes |
| security.gdpr | yes | - |

## Capabilities (Llm Gateways)

| Capability | Kong AI Gateway | LiteLLM |
|---|:--:|:--:|
| **Capabilities** |  |  |
| Openai compatible API | ✓ | ✓ |
| Automatic failover | ✓ | ✓ |
| Routing strategy | - | Cost based |
| Caching | Semantic | - |
| Virtual key management | ✓ | ✓ |
| Budget spend controls | ✓ | ✓ |
| Rate limiting | ✓ | ✓ |
| Guardrails moderation | - | Built in |
| Request cost observability | ✓ | ✓ |
| Streaming support | - | - |
| Deployment model | - | Self hosted |
| Open source | ✓ | ✓ |
| Pricing model | - | Open source self host |

*Source: Vioscale. Generated 2026-09-01T17:01:14.396Z. "-" = undocumented, not absent.*