Developers Criticize Inkling 2 mini's Memory-Hungry Long Context
Developer gnukeith, who is fine-tuning Thinking Machines' Inkling 2 mini for a small information-gathering agent, says its long-context runs consume excessive memory, forcing frequent compaction and squeezing out room for running multiple agents in parallel—calling it the model's biggest flaw.
2026-11-22 ~ 2026-11-22 · 2 related posts
- Dev calls out Inkling 2 mini: 34GB RAM for 1M context vs DeepSeek Flash's <1GB — gnukeith · 2026-11-22
- Long context is Thinking Machines' biggest flaw, dev says in follow-up — gnukeith · 2026-11-22