Claim◆Article

Inkling is not the strongest overall model available today, open or closed; instead its multimodal capabilities, efficient thinking, and availability on Tinker make it a good open-weights base for customization via fine-tuning.

Thinking Machines Lab frames Inkling not as a state-of-the-art model but as a strong, efficient base meant to be fine-tuned through their Tinker platform. ✦ AI generated

Thinking Machines Lab · Simon Willison's Weblog · 2026-07-16 · original ↗

Share card ↓Export .md Export .txt

Inkling is not the strongest overall model available today, open or closed. Instead, a combination of qualities makes it a good open-weights base for customization: multimodal capabilities, efficient thinking, and availability on Tinker for fine-tuning.

Read full article ↗excerpt · fair-use quotation

▦ Thinking Machines Lab ✳ Tinker

Related moments

Mechanism◆Video · 32:53 · 2m

Using accumulated human tax-expert feedback on model errors as ground truth and reward signal made OpenAI's reinforcement fine-tuning (RFT) especially effective for Sphere's tax-determination task, since it targeted precisely the hard cases the model had previously missed.

Alex Boucott · The TWIML AI Podcast

Mechanism◆Video · 35:48 · 2m

Founders shouldn't just tell their own inward-facing company story; instead they should identify the most interesting things happening in the world and explain how their company relates to them.

Marc Andreessen · a16z Podcast

Data◆Video · 43:13 · 2m

Kumo's relational foundation model beats the best supervised models on benchmark tasks by about 5% relative accuracy out of the box, rising to a 12% improvement after fine-tuning, gains large enough to translate into tens of millions of dollars of business impact in production systems.

Jure Leskovec · The TWIML AI Podcast