Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0487
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-07-22T03:56:56.663110Z
status: published
lens: trace
sector: science
headline: The Illusion of Thinking
dek: Recent generations of frontier language models have introduced Large Reasoning Models (LRMs) that generate detailed thinking processes before providing answers. While these models demonstrate improved performance on reasoning benchmarks, their fundamental capabilities, scaling properties, and limitations remain insufficiently understood. Current evaluations primarily focus on established mathematical and coding benchmarks, emphasizing final answer accuracy. However, this evaluation paradigm often suffers from da…
gain_title: Large Reasoning Models generate detailed thinking traces before answering and demonstrate improved performance on reasoning benchmarks, with advantage over standard LLMs on medium-complexity controllable puzzles.
problem_title: Frontier Large Reasoning Models face a complete accuracy collapse beyond certain puzzle complexities and exhibit a counterintuitive scaling limit where reasoning effort declines despite adequate token budget.
trace_subject: performance of Large Reasoning Models on controllable puzzles of varying compositional complexity
gain_reading: Large Reasoning Models generate detailed thinking traces before answering and demonstrate improved performance on reasoning benchmarks, with advantage over standard LLMs on medium-complexity controllable puzzles.
gain_evidence: generate detailed thinking processes before providing answers | demonstrate improved performance on reasoning benchmarks
problem_reading: Frontier Large Reasoning Models face a complete accuracy collapse beyond certain puzzle complexities and exhibit a counterintuitive scaling limit where reasoning effort declines despite adequate token budget.
problem_evidence: face a complete accuracy collapse beyond certain complexities | reasoning effort increases with problem complexity up to a point, then declines despite having an adequate token budget
quick_read: Published September 23, 2025, this peer-reviewed study systematically tested frontier Large Reasoning Models that generate detailed thinking processes before answering. Using controllable puzzle environments to vary compositional complexity, the authors analyzed final accuracy and internal reasoning traces and compared LRMs to standard LLMs under equivalent inference compute.

The work matters because it moves evaluation beyond final-answer accuracy on contaminated math and coding benchmarks to trace structure and scaling behavior. It remains uncertain how these puzzle-based collapse patterns and inconsistent algorithm use generalize to real-world scientific or applied reasoning workflows outside the controlled environments.
limitation: Findings are bounded to controllable puzzle environments and reveal that LRMs have limitations in exact computation and inconsistent reasoning across puzzles.
tag: Automated dual reading
key_points: Study uses controllable puzzle environments that allow precise manipulation of compositional complexity while maintaining consistent logical structures. | Comparison under equivalent inference compute identifies three regimes: low-complexity where standard models outperform LRMs, medium-complexity where LRMs have advantage, high-complexity where both collapse. | Analysis finds LRMs fail to use explicit algorithms and reason inconsistently across puzzles, with limitations in exact computation.
rundown: The authors evaluated frontier LRMs using controllable puzzle environments that enable precise manipulation of compositional complexity while keeping logical structures consistent, allowing inspection of both final answers and internal reasoning traces.

Across puzzles they observed three regimes under equivalent inference compute: standard LLMs outperforming LRMs at low complexity, LRMs advantaged at medium complexity, and both collapsing at high complexity, with LRMs failing to use explicit algorithms and showing inconsistent reasoning.
sources:
- peer_reviewed | SuperIntelligence - Robotics - Safety & Alignment | https://doi.org/10.70777/si.v2i6.15919 | 2025-09-23
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

a56d29a25f21da808e2e0d2a921df2f2b127817a5512d1a0003b6520b011ed47

previous

0000000000000000000000000000000000000000000000000000000000000000