🛰️ Daily AI Frontier
‹ back to 2026-08-13

Who Thinks Best Depends on How Long You Let Them: Budget-Dependent Rankings in LLM Evaluation

Research LLMs & Foundation Models

Ranking

Overall 84
Content 95
Popularity 58

Observed public metrics from 1 member.

Merged summary

TL;DR - LLM rankings can change or reverse as generation-token budgets vary, undermining evaluations performed at a single inference budget. Budget-conditioned evaluation and model routing may better reflect accuracy, efficiency, and model complementarity.

  • Across 56,476 inferences, 3–19% of items became less accurate with larger budgets.
  • Model rankings reversed across budgets on all three reasoning benchmarks.
  • Oracle model selection improved performance by up to 27.8 percentage points, especially under constrained budgets.
  • A budget-aware router captured 14.1% of the cross-domain oracle gap, but budget features transferred poorly between domains.

Sources (1)

Who Thinks Best Depends on How Long You Let Them: Budget-Dependent Rankings in LLM Evaluation

arXiv cs.AI Rodrigo Guedes de Souza, Alison R. Panisson 2026-08-12 arXiv:2608.12150
Public signals Semantic Scholar citations 1 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 1 · Influential citations 0 X · N/A Fetched 2026-09-12 14:28:10.194994 UTC

TL;DR - LLM rankings can change or reverse as generation-token budgets vary, undermining evaluations performed at a single inference budget. Budget-conditioned evaluation and model routing may better reflect accuracy, efficiency, and model complementarity.

  • Across 56,476 inferences, 3–19% of items became less accurate with larger budgets.
  • Model rankings reversed across budgets on all three reasoning benchmarks.
  • Oracle model selection improved performance by up to 27.8 percentage points, especially under constrained budgets.
  • A budget-aware router captured 14.1% of the cross-domain oracle gap, but budget features transferred poorly between domains.
item →