Chutes Sees TPS Growth, Deployment Issues Reported

Chutes models including Qwen 3.5-397B and GLM 5.1 achieved higher throughput and paygo revenue gains over the past week. Team identified deployment bottlenecks: image builds stalling and voxcpm loader errors blocking chute instantiation. Inventory visibility and GPU idle-capacity strategy flagged as near-term priorities.

• •95 TPS on Qwen 3.5-397B-A17B-TEE; GLM 5.1 and Kimi K2.6 also climbing
• •Deployment failures: image builds pending indefinitely, voxcpm InductorError on startup
• •Team planning more reliable inventory UI to expose available capacity