Run every model as one system.
Every model. One API. Any cloud.
Control what runs. Govern every request.
Your models span providers, keys, and clouds.Inferbase gives you one place to control them all.
The right model for every request.
Set model: auto. Every prompt is matched to capable models, then routed to the best fit for your objective.
Paste your prompt below, or try it in the Playground.
The real classifier and engine over the managed pool, no model called. In production the same decision arrives before the first token and is kept by request id.
Run it in the Playground →Your rules on every request.
Which models may serve, what guardrails check, which tools an agent may call, and what each project may spend.
See the policies- model
- autoPer request
- objective
- balancedPer request
- quality preset
- standard
- scope
- managed catalog + your keys
- model pool
- any qualifying model
- session affinity
- onPer request
Changed in the dashboard or the API; the next request uses the new settings, no deploy.
Every request, accounted for.
Which model answered and why, what it cost and how long it took, kept for every request.
See what is recordedTotal requests
48,210
Smart-routed
41,876
Direct
6,334
Total tokens
61.4M
Requests over time
Smart-routedDirect
Requests by model
- DeepSeek V4 Flash41%
- GPT-OSS 120B27%
- GLM 4.7 Flash18%
- Other models14%
| Source | Model | Tokens | Cost | TTFT | Latency | Time | |
|---|---|---|---|---|---|---|---|
| prod-api | DeepSeek V4 Flashauto | 508 | $0.0001 | 417 ms | 1.2 s | 14:02:11 | |
| prod-api | GPT-OSS 120Bauto | 2,914 | $0.0009 | 380 ms | 3.4 s | 14:02:09 | |
| Playground | GLM 4.7 Flash | 1,206 | $0.0004 | 512 ms | 2.1 s | 14:01:58 |
Put intelligence in the middle.
base_url = "https://api.inferbase.ai/v1" model = "auto"
Keep your OpenAI or Anthropic SDK. Free to start, with 5,000 requests a month and no card.
How Inferbase compares to OpenRouter, LiteLLM, Portkey and others →