Do reasoning models actually beat standard models on optimization?
Explores whether extended chain-of-thought in reasoning models delivers performance gains on constraint-satisfaction problems like power-grid optimization. Matters because reasoning models are treated as automatic upgrades, but the evidence may not support that claim.
Reasoning models have been treated as a generalized capability upgrade — more thinking tokens at test time, broadly better performance. On constraint-bound numerical optimization the upgrade does not materialize. Reasoning variants do not systematically outperform their non-reasoning counterparts on power-grid, financial-operations, or cyber-security feasibility problems. The longer trace does not become a longer iteration.
The reason this matters: extended chain-of-thought looks like it should help. The problem involves multi-step arithmetic, interacting constraints, and convergence-style reasoning — exactly the regime where "think more" is supposed to pay. The data say it does not. Whatever extended CoT is doing on these tasks, it is not running a Newton-Raphson iteration or a primal-dual update in latent space; it is producing more text without producing more computation.
This is consistent with a growing view that reasoning models excel where the bottleneck is exploration over reasoning paths (math contests, code, multi-hop QA) but stall where the bottleneck is numeric procedure. Constraint satisfaction over real physical systems is the latter. Adding chain length adds search over verbal restatements of the problem, not iterations of the algorithm that would solve it.
The implication for product: choosing "reasoning model" for an optimization-heavy workflow is not automatically the right call. The relevant decision is whether the bottleneck is verbal reasoning or numeric computation. If numeric, the cost-effective path is hand-off to a solver, not more thinking tokens.
Inquiring lines that read this note 72
This note is a source for these research framings, grouped by the broader line of inquiry each explores. Scan the bold lines of inquiry; follow any specific question forward.
How does model scale affect fine-tuning and downstream task performance?- Why do only two of fourteen models improve when problem constraints are removed?
- What production constraints should determine paradigm selection?
- Why do production systems optimize for three model classes instead of foundation models?
- What makes frozen model reasoning different from weight-based parameter updates?
- When does the right constraint beat additional model capacity?
- Can weaker models match stronger ones with sufficient search and reasoning budget?
- How do KV cache pruning and subproblem contraction both free reasoning capacity?
- Can reasoning models outperform non-reasoning models with more inference compute?
- How do unstated feasibility constraints affect model decision-making?
- Can explicit constraint statements override the dominance of surface heuristics?
- What distinguishes intrinsic search from extrinsic search method approaches?
- What design changes could make constraint inference more reliable without explicit cuing?
- Do tool-enabled reasoning models close the gap on constraint satisfaction?
- Why do non-reasoning models work better under extreme decomposition than reasoning models?
- Can explicit optimal algorithms prevent reasoning model collapse at high complexity?
- Is the reasoning cliff actually a tool-use problem?
- What makes constraint satisfaction problems epistemically cleaner than other reasoning tasks?
- Which constraint types do reasoning models handle best?
- What limits external scaling when a model lacks reasoning foundation?
- Do reasoning-enabled and prompt-hardened conditions show the same architectural penalty?
- How does nesting optimization levels improve on traditional network depth?
- What makes deterministic recursive reasoning models underperform on multi-solution tasks?
- Does iterative computation for reasoning transfer to environment dynamics modeling?
- Why does the second loop do most of the productive refinement work?
- Why does reapplying the same computation stages improve model performance?
- Do iterative refinement methods reproduce the same overthinking failure mode?
- Can chain-of-thought explanations be both sufficient and necessary for model decisions?
- How much does chain-of-thought reasoning actually determine model outputs?
- Can external verifiers replace reasoning trace quality in solution guarantees?
- What separates verifiable reasoning from open-ended judgment in scaling requirements?
- Why does most refinement in iterative models maintain answers rather than improve them?
- How does symbolic solver feedback differ from language-based self-critique?
- Does architectural design matter more than model scale for reasoning tasks?
- What makes multi-paradigm chaining a distinct reasoning topology?
- Why does augmenting symbolic reasoning outperform replacing it entirely?
- Can static reasoning patterns work better than dynamic branch selection?
- Can parallel reasoning chains outperform longer sequential chains with the same compute?
- When does sequential reasoning provide exponential advantages over parallel voting?
- When does sequential chain-of-thought dramatically beat parallel voting approaches?
- Can optimization algorithms exploit the shift between procedural and planning bottlenecks?
- Can the LLM-Modulo framework extend solver integration to domain planning?
- Can symbolic solvers reliably replace LLM reasoning for logical tasks?
- Do higher asymptote recipes unlock genuinely novel reasoning strategies?
- Can reasoning improvements be attributed when optimizer and scaffold are unknown?
- When is detailed step-by-step reasoning actually counterproductive for solving a problem?
- Why does extended chain-of-thought reasoning fail to improve numerical optimization performance?
- Why do reasoning models fail to improve constrained optimization performance?
- How does making implicit reasoning requirements explicit change model performance?
- How should benchmarks evaluate workflow architecture versus raw model performance?
- Do reasoning benchmarks predict real performance in long delegated workflows?
- Why do model evaluation and downstream evaluation require different benchmark designs?
- What planning strategies reduce execution steps without sacrificing solution quality?
- How does planning-before-execution compare to iterative reasoning and action loops?
Related concepts in this collection 4
This note in its neighbourhood — explore the map, then jump to a related concept in the list below.
Click a node to walk · click center to open · click Open in graph to see this note in the full knowledge graph
-
Do larger language models solve constrained optimization better?
Explores whether scaling LLMs—through more parameters, better training, or reasoning extensions—improves their ability to satisfy constraints in real optimization problems like power grids and portfolios.
same paper, the parent finding
-
Do large language models actually perform iterative optimization?
Explores whether LLMs execute genuine numerical procedures like Newton-Raphson or instead pattern-match to memorized solution templates when solving constrained optimization problems.
same paper, the mechanism
-
Does more thinking time always improve reasoning accuracy?
Explores whether extending a model's thinking tokens linearly improves performance, or if there's a point beyond which additional reasoning becomes counterproductive.
adjacent: CoT ceilings in general
-
Why does chain of thought accuracy eventually decline with length?
Explores why longer reasoning chains don't always improve answers, and how the optimal length shifts based on task difficulty and model capability.
adjacent: more thinking is not monotonically better
Related papers in this collection 8
Papers most semantically related to this note, ranked by cosine similarity in the embedding space.
- Can Large Language Models Reason and Optimize Under Constraints?
- The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
- Chain of Thoughtlessness? An Analysis of CoT in Planning
- Comment on The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
- On the Reasoning Capacity of AI Models and How to Quantify It
- Evaluating Theory of Mind in Reasoning Models: Robustness over Reasoning
- A Comment On "The Illusion of Thinking": Reframing the Reasoning Cliff as an Agentic Gap
- Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
Original note title
reasoning models do not systematically outperform non-reasoning models on real numerical optimization — extended chain-of-thought is not a substitute for iterative computation