Tests find GPT-6.1 Sol's allowance is ~20% lower than 6 Sol, fueling claims of intentional throttling
BorisMPower · x · 2026-10-01
Blogger Fei2411 benchmarked GPT-6.1 Sol with a uniform workload and got conclusions opposite to community lore:
- The model is indeed slow (community tests hit as low as 10+ tok/s), but its allowance is not generous — measured 20% lower than 6 Sol, which launched only 6 days ago.
- The Pro 200 tier looks even worse value: its multiplier vs GPT Plus was cut, and GPT Plus itself is shrinking too.
- The author suspects OpenAI of intentional throttling to create the illusion of a large allowance via slow consumption.
The author also open-sourced real-api-pricing: it computes each subscription's real per-million-token price (monthly fee ÷ usable tokens) and plots allowance, unit price, and Pareto charts, already at 672 stars.
Related event: GPT-6.1 Sol Quota Reportedly 20% Lower Than 6 Sol in Tests(2 posts)→
More from Models
- Initial probing of Opus suggests its time estimates are hallucinated from human training samples — StefanoGogioso · 2026-10-01
- Arena benchmarks Claude Sonnet 5.5 vs GPT-6.1 Sol: can cheap tiers match flagship results? — arena · 2026-10-01
- Inception launches Mercury Decide on OpenRouter: a decision model doing 14 decisions/sec — StefanoErmon · 2026-10-01
- GPT-6.1 Sol beats Astra on low-resource translation, dev benchmark shows — nickbaumann_ · 2026-10-01
- ChatGPT $200 tier subscriber calls expiring compensation credits 'a joke' — mike64_t · 2026-10-01
- Open-Source Models Like GLM 5.3 Flash and DeepSeek V4.1 Flash Now Outpace Last Year's Frontier — victormustar · 2026-10-01