GPT-6.1 Sol allowance quietly cut ~20% vs 6 Sol, possibly masked by throttling
RexDouglass · x · 2026-10-01
@Fei2411's controlled tests found GPT 6.1 Sol's subscription allowance is nearly 20% lower than 6 Sol's, released less than a week earlier.
- Uniform workloads rule out cache-hit-rate differences.
- The community's feeling that "6.1 Sol burns allowance slowly" is real — because the model itself is extremely slow, measured as low as 10+ tok/s.
- Author suspects OpenAI deliberately throttles to create an illusion of generous allowance.
- Conclusion: the $200 Pro plan is even worse value: the multiplier vs GPT Plus was cut, and Plus allowances are shrinking too.
Accompanying open-source project real-api-pricing (672 GitHub stars): computes real per-token cost as monthly fee ÷ usable tokens for AI coding subscriptions, plotting price vs public leaderboard scores in interactive Pareto charts.
Related event: GPT-6.1 Sol Quota Reportedly 20% Lower Than 6 Sol in Tests(2 posts)→
More from Models
- Stealth Model 'silicon-2' Surfaces in Devin, Suspected to Be Anthropic's Next Model — kimmonismus · 2026-11-16
- Google hides Gemini model label in AI Mode, rolls Gemini 4 Carbon to Pro subscribers — gaganghotra_ · 2026-11-16
- Open-weight Karpo-Decision-31B tops Decision Index at 73.9, one forward pass per answer — CodeByPoonam · 2026-11-16
- Nathan Lambert calls out Ethan Mollick for 'selectively ignoring' Kimi in model comparisons — natolambert · 2026-11-16
- Eval Firm Trajectory Told 40 Agents "NEVER Stop" in Harness Without Telling Epoch — peterwildeford · 2026-11-16
- Unsloth Q8 crushes default Q3 in a self-hosted test over 331 financial PDFs — No_Cartographer3953 · 2026-11-16