Merge Gateway slashes GLM 5.3 Flash pricing 90%

Sub-cent execution for benchmark-competitive models highlights how aggressive inference pricing and gateway routing are eroding proprietary model margins.

• –Aggressive price-performance wedge: Slashing task costs to $0.009 while maintaining benchmark parity with models costing 31 cents or more creates a compelling economic incentive for developers to swap foundation models.
• –Gateway as an acquisition layer: By subsidizing GLM 5.3 Flash, Merge leverages loss-leader model pricing to drive developer adoption of its unified API and governance platform.
• –Open-weight pressure on closed APIs: As hybrid attention open-weights models achieve frontier-tier results at fractional costs, proprietary API providers will struggle to justify 30x price premiums on mid-tier workloads.
• –Narrow evaluation window: The time-limited discount running through September 30 gives engineering teams a short window to benchmark high-throughput agent and coding workflows at near-zero inference cost.