Mira Murati's Thinking Machines Lab has released Inkling, a 975B-parameter model that demonstrated self-fine-tuning capabilities on its launch day. The model successfully trained itself to become a lipogram model, avoiding the letter 'e' through a seven-step post-training process. On release, Inkling achieved a score of 41 on the Artificial Analysis Intelligence Index, surpassing NVIDIA's Nemotron 3 Ultra and other open-weight models. Despite its strong performance, the lab acknowledged that Inkling is not the strongest model available overall. AI
IMPACT Demonstrates advanced self-improvement capabilities in LLMs, potentially accelerating model development and specialization.
RANK_REASON Frontier lab model release with novel self-fine-tuning capability and benchmark performance. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Artificial Analysis
- Gemma 4.31B
- GPT-OSS 120B
- Inkling
- Mira Murati
- Nemotron 3 Ultra
- NVIDIA
- Thinking Machines Lab
- Tinker API
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
