Thinking Machines Releases Inkling 2 Small, a 594B-Parameter Multimodal MoE
[AINews] Inkling 2 Small: 17A-594B trained in 68 hours on 12T tokens, beating Muse Spark
What happened
Thinking Machines released Inkling 2 Small, a multimodal mixture-of-experts model with 594B total parameters and 17B active parameters, available for inference and fine-tuning on Tinker. The roundup reports 68 hours of pretraining on GB300s and an Artificial Analysis Intelligence Index score of 31, while noting that Apache 2.0 weights are promised for next week after external safety testing. Independent evaluations also report high output-token use per task.
Why it matters
The roundup compares Inkling 2 Small across independent benchmarks and partner evaluations, while noting its high output-token use and that open weights are still pending safety testing.
Read next
More changes in today’s briefingMETR Says Anthropic May Have Crossed Its AI Acceleration Warning Threshold in JulyKeep up with AI, one email a day
Model releases, price changes and compute developments, with original sources. Free, confirmed by email, and easy to unsubscribe. Read a past issue · 2026-11-07