Skip to content
NewTry RouterPlus with free creditsTry playground →
← All posts · News

Inkling 2 Small launches on Tinker at $0.30/$1.00, weights to follow

Thinking Machines' second-generation model: 594B MoE, 17B active, closed weights for now.

Thinking Machines Lab released Inkling 2 Small on November 5, 2026, the first model of its second Inkling generation. It is available now on Thinking Machines' Tinker platform for inference and fine-tuning, with closed weights at launch. The lab says it intends to release the full weights under Apache 2.0 next week, after external safety testing and red-teaming.

What changed

  • Mixture-of-Experts model with 594B total parameters and 17B active per token. Text and image input, text output, with variable effort. No audio input, unlike the first-generation Inkling models.
  • Tinker ID: thinkingmachines/Inkling-2-Small. The model card lists a context window of up to 1M tokens; Tinker's serverless endpoint is listed at 256K.
  • Tinker serverless inference is in beta, with a waitlist for production use.
  • Thinking Machines says the model will reach additional platforms in the coming weeks, after safety testing. Checkpoints become downloadable when the weights are released.
  • The lab notes the model is less optimized for chat than for agentic and coding work.

Prices

Per million tokens. Tinker prices are from Thinking Machines' Tinker pricing page; the RouterPlus row is from the RouterPlus catalog on November 7, 2026.

ModelWhereInputCached inputOutput
Inkling 2 SmallTinker serverless (256K)$0.30$0.03$1.00
Inkling-SmallTinker serverless (256K)$0.30$0.06$1.20
Inkling SmallRouterPlus via DeepInfra$0.45$0.10$1.20

Fine-tuning Inkling 2 Small on Tinker costs $1.50 per million training tokens at 64K context and $3.00 at 256K.

What it means

Inkling 2 Small is not in the RouterPlus catalog as of November 7, 2026. The first-generation thinkingmachines/inkling-small is, through DeepInfra, at api.routerplus.com/v1. Teams that want to self-host or move to another host should wait for the Apache 2.0 weights and check the licence file when it lands.

Sources

  1. Thinking Machines Lab: Introducing Inkling 2 Small (November 5, 2026)
  2. Inkling 2 Small model card
  3. Tinker docs: Models & Pricing
  4. RouterPlus model catalog