How we compare, row by row, with the source on every row
Hanzo vs
Each page sets a Hanzo product beside a named alternative.
| Row | EnsoHanzo | OpenRouter | Portkey | LiteLLM |
|---|---|---|---|---|
| 600+ | 500+ | 1,600+ | 1,800+ | |
| Yes | Yes | Yes | Yes | |
| Yes | Yes | Yes | Yes | |
| Yes | No | No | No | |
| The router beside single models from other labs on the same 198 questions, every row a research run with its accuracy and what its tokens cost at list prices | not measured by us | not measured by us | not measured by us | |
| Yes | No | Not published | Yes | |
| Yes | No | Self-hosting offered at enterprise tier | Yes |
- Enso
- 600+
- OpenRouter
- 500+
- Portkey
- 1,600+
- LiteLLM
- 1,800+
Enso: 600+. OpenRouter: 500+ (80+ providers). Portkey: 1,600+. LiteLLM: 1,800+ (140+ providers).
- Enso
- Yes
- OpenRouter
- Yes
- Portkey
- Yes
- LiteLLM
- Yes
Enso: Yes — /v1/chat/completions. OpenRouter: Yes. Portkey: Yes. LiteLLM: Yes.
- Enso
- Yes
- OpenRouter
- Yes
- Portkey
- Yes
- LiteLLM
- Yes
Enso: Yes — same key. OpenRouter: Yes — Anthropic Messages format, tools and extended thinking. Portkey: Yes — /v1/messages routes to any of their models. LiteLLM: Yes — serves the format at /v1/messages.
- Enso
- Yes
- OpenRouter
- No
- Portkey
- No
- LiteLLM
- No
Enso: Yes — model calls are quoted from above before they happen, and an agent that would breach its cap is refused rather than discovered on an invoice. OpenRouter: No — blocks once a limit is reached (403), Enterprise plan only. No pre-flight quote published.. Portkey: No — budget limit expires the key once reached, Enterprise plan. No pre-flight quote published.. LiteLLM: No — caps budgets and stops at the cap. No pre-flight quote published..
- Enso
- The router beside single models from other labs on the same 198 questions, every row a research run with its accuracy and what its tokens cost at list prices
- OpenRouter
- not measured by us
- Portkey
- not measured by us
- LiteLLM
- not measured by us
Enso: The router beside single models from other labs on the same 198 questions, every row a research run with its accuracy and what its tokens cost at list prices. OpenRouter: not measured by us. Portkey: not measured by us. LiteLLM: not measured by us.
- Enso
- Yes
- OpenRouter
- No
- Portkey
- Not published
- LiteLLM
- Yes
Enso: Yes — GPQA-Diamond, 198 graduate-level questions, every row a research run. OpenRouter: No — their Benchmarks API scores the models, not the router. Portkey: Not published. LiteLLM: Yes — Auto Routing Cost Ladders: 58.2% savings, 46.8% win rate vs flagship.
- Enso
- Yes
- OpenRouter
- No
- Portkey
- Self-hosting offered at enterprise tier
- LiteLLM
- Yes
Enso: Yes — the whole cloud is one binary and one API, on your own Kubernetes or your laptop. OpenRouter: No — a hosted service. Portkey: Self-hosting offered at enterprise tier. LiteLLM: Yes — open source, self-hosted free forever.
Count the Yes
- EnsoHanzo5 / 5
- OpenRouter2 / 5
- Portkey2 / 5
- LiteLLM4 / 5
A different question
If what you are weighing is whether to run it yourself rather than buy it, that page is separate.
Model routing APIs