Users Claim Codex Subscription Throttled to a Third of API Speed

User benchmarks claim Codex subscribers get a throttled version of the same model at roughly 22 tokens/s, about a third of API speed, with some suggesting OpenAI slowed inference to conserve usage quotas. The findings are unverified user-side measurements.

2026-10-01 ~ 2026-10-01 · 2 related posts

• Codex subscribers allegedly getting throttled models: 22 t/s vs 73 t/s via API — RexDouglass · 2026-10-01
• User benchmarks claim OpenAI quietly cut model speeds to ~60% to stretch usage limits — bdsqlsz · 2026-10-01