Inference platform from Keln Inc. (San Francisco) that routes open-weight models across independent compute providers at one fixed per-token price per model, with a cheaper Lite service tier, OpenAI- and Anthropic-compatible APIs and agent sandboxes. Card top-ups via Stripe.

Features

OpenAI Compatible
Streaming
Batching
Fine-tuning
Embeddings
Vision
Audio
Function Calling
JSON Mode

Compliance

None listed

Pricing Model

per token

Details

Models 18
API Base https://api.keln.ai/v1
EmbeddingsLLM Inference

Model Catalog (18)

Model Type Input $/1M Output $/1M Context Speed Status
DeepSeek V4 Pro
DeepSeek
llm $1.32 $3.96 — —
DeepSeek V4 Pro
DeepSeek
Lite
llm $0.924 $2.77 — —
DeepSeek V4.1 Flash
DeepSeek
llm $0.300 $1.20 — —
DeepSeek V4.1 Flash
DeepSeek
Lite
llm $0.210 $0.840 — —
GLM 5.2
Z.ai
llm $1.40 $4.40 — —
GLM 5.2
Z.ai
Lite
llm $0.770 $2.42 — —
GLM 5.3
Z.ai
llm $1.40 $4.40 — —
GLM 5.3
Z.ai
Lite
llm $0.910 $2.86 — —
GLM 5.3 Flash
Z.ai
llm $0.150 $0.500 — —
GPT OSS 120B
OpenAI · 117B MoE (5.1B active)
llm $0.150 $0.600 — —
GPT OSS 120B
OpenAI · 117B MoE (5.1B active)
Lite
llm $0.075 $0.300 — —
Gemma 4 31B
Google · 27B
llm $0.150 $0.460 — —
Kimi K3.1 Pro
Moonshot AI
llm $2.00 $10.00 — —
MiniMax M3.1 Flash
MiniMax
llm $0.150 $0.600 — —
Qwen3 Embedding 8B
Alibaba · 8B
embedding $0.013 — — —
Qwen3.8 2.4T A95B
Alibaba
llm $2.50 $6.25 — —
Qwen3.8 27B
Alibaba
llm $0.500 $3.20 — —
Qwen3.8 27B
Alibaba
Lite
llm $0.350 $2.24 — —

Keln API FAQ

Is the Keln API free?

Not as far as we track: Keln has no free API tier on record. The cheapest option we track is Qwen3 Embedding 8B at $0.01 in per 1M tokens.

What is the cheapest Keln model?

Qwen3 Embedding 8B at $0.01 in per 1M tokens, based on the 18 models we track for Keln.

How many models does Keln offer?

We track 18 Keln models: 17 LLM, 1 embedding.

Is the Keln API OpenAI-compatible?

Yes. Keln exposes an OpenAI-compatible endpoint, so you can use the OpenAI SDKs by changing the base URL and API key.

Subscribe to our newsletter

News and updates, 1–2 emails per month. No spam.