Thinking Machines Releases Inkling-Small

Thinking Machines just shipped Inkling-Small: 276B total params, only 12B active via MoE, but keeps the 1M-token context and variable thinking effort from the full Inkling model. Same reasoning capability class, far less compute needed to run it.

Open original source ↗Generate a post from this angle

For writing research only. Verify the original source before publishing; market data is not investment advice.