Thinking Machines Lab

thinkingmachines/…

Get an API key

Every Thinking Machines Lab model we serve, with the numbers most providers leave out — the quantisation each endpoint actually runs at, measured first-token latency, and real uptime. Model ids are copyable: pass one straight to any OpenAI-compatible client.

Inkling Small

thinkingmachines/inkling-small

·Thinking Machines Lab

Natively multimodal 276B mixture-of-experts model from Thinking Machines Lab, activating roughly 12B parameters per token. Text, image and audio in; text out. Strong on agentic and long-horizon work.

Parameters

276B (12B active)

Max context

512K

Quantisation	Context	p50 TTFT	Throughput	Uptime	Input /M	Output /M	Cached in /M	Status
Undisclosed	512K	851ms	—	99.96%	$0.45	$1.20	$0.10	live

moelong-contextmultimodalvisionaudioevaluation

thinkingmachines/Inkling-Small ↗Details →

Other labs we serve

Qwen6DeepSeek3Ornith3Zz-ai3Infersia2Google1Mmeta-llama1Meta1Mminimax1OpenAI1Ttencent1xAI1Xxiaomi1ZZeroEntropy1