Price & performance

Cheapest per Token providers

Every model provider on LLMAPI tagged “Cheapest per Token” — 11 of 50 providers. All of them run through one API key and one OpenAI-compatible endpoint.

Providers listed11
Models reachable53
Filters available84
IntegrationOne OpenAI-compatible endpoint

11 providers matching this filter.

ProviderModelsRegionsBest for
Google

Gemini models via AI Studio and Vertex AI.

7 Global Multimodal workloads and very long context at aggressive price points View →
Mistral

Efficient open and commercial models from Paris.

10 EU, US Price-performance and EU data-residency requirements View →
DeepSeek

Open reasoning models at aggressive price points.

5 Global Reasoning and coding at the lowest per-token prices View →
Zai (Zhipu AI)

GLM series models from Zhipu AI.

5 APAC Chinese-English bilingual applications View →
Multiverse Computing

Compressed models via quantum-inspired techniques.

3 EU Compressed models for on-device and cost-sensitive deployments View →
Parasail

Serverless inference marketplace.

4 US Serverless open-model inference at spot-style pricing View →
Crusoe

Energy-first GPU cloud.

1 US Cost-effective, sustainably powered GPU compute View →
DeepInfra

Low-cost inference for popular open models.

8 US Popular open models at the lowest per-token cost View →
Novita

Affordable GPU cloud and model APIs.

4 APAC, US Budget-friendly serverless GPUs and model APIs View →
Wafer

Distributed inference network.

3 Global Affordable open-model inference on distributed GPUs View →
Makora

Model hosting and inference APIs.

3 Global Pay-as-you-go hosting for open models View →

One key, every provider on this page

Consolidated billing, spend limits and instant fallback between providers.