Tinker Slashes AI Prices Up to 70% Following Efficiency Gains

Tinker reduced its API pricing by up to 70% following efficiency improvements for long-context reinforcement learning. The company launched GLM-5.3-Flash and DeepSeek-v4.1-Flash, with the GLM-5.3-Flash model costing 4 to 5 times less than the standard GLM-5.3 version.