Thinking Machines has released Inkling 2 Small, a 594 billion parameter model that achieves frontier-level performance on benchmarks. The model utilizes a mixture-of-experts architecture, with only 17 billion parameters active per token, allowing it to compute efficiently while retaining extensive knowledge. While currently only accessible via the Thinking Machines platform, Tinker, with competitive pricing, the lab plans to release the full weights under an Apache 2.0 license next week, potentially making it the most powerful open-weight model available. AI
IMPACT Potentially democratizes access to a frontier-level model, driving innovation in open-source AI development.
RANK_REASON Frontier-lab model release with system card and planned open weights. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Apache Software License 2.0
- Artificial Analysis Intelligence Index
- Gemini 3.8 Flash
- GPQA Diamond
- Inkling 2 Small
- Kimi K3.1 Pro
- Mira Murati
- NVIDIA GB300 NVL72
- SWE-Bench Pro v2
- SWE-bench Verified
- Thinking Machines
- Tinker
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →