SYNTHESIS NOTE
Topics›Test Time Compute›this note

Can non-reasoning models catch up with more compute?

Explores whether inference-time compute budget can close the performance gap between standard models and those trained for reasoning, and what training mechanisms might enable this.

Synthesis note · 2026-02-20 · sourced from Test Time Compute

In verifier-free inference-time compute experiments (Think Deep, Think Fast), non-reasoning models fall substantially behind reasoning models even when given an extremely high inference budget. The gap doesn't close with more compute — it just stays there.

This sets a hard limit on Can inference compute replace scaling up model size?. The substitution works within a training regime, but not across training regimes. A standard instruction-tuned model with more inference compute cannot replicate what a model trained specifically for extended reasoning can do, even given equivalent token budgets.

Why? Reasoning models have internalized the reasoning process through training — they know how to use additional tokens productively. Non-reasoning models don't have this structure, so additional tokens degrade into noise or verbosity rather than improved reasoning. The training regime instills the reasoning protocol that makes inference compute usable.

Qualification from targeted activation (Base Models paper): The gap is substantially closeable through targeted steering of base model activations without weight updates. A hybrid model using base model weights + thinking model deployment decisions recovers 91% of the performance gap while steering only 12% of tokens. This doesn't invalidate the finding — non-reasoning models without steering still fall behind — but it significantly changes what "non-reasoning model" means in practice. If capability already exists latently and steering can surface it, the gap is about deployment mechanisms, not raw capability. See Does RL teach reasoning or just when to use it?.

The imitation learning ceiling (Tutorial on LLM Reasoning): SFT/imitation learning creates an intelligence upper bound: the model is bounded by the quality of demonstrations it learns from, unable to surpass the skill level present in training data. RL + world models is the path beyond this ceiling, because RL allows discovery of strategies that exceed any individual demonstration. This provides the mechanism for why reasoning-specific training matters: it is not merely "more training" but training that enables exceeding the imitation ceiling.

This is a strong argument for the necessity of reasoning-specific post-training, not just inference-time tricks. Compute can amplify capability but cannot manufacture it. The dependency on training regime appears to be capability-specific: Can language models learn grammar from child-scale data? — syntactic competence scales down readily, achievable with human-scale data and the right composition. Reasoning capability requires the opposite: specialized training that instills the reasoning protocol itself. The lesson is not "you need a bigger model" but "you need the right training for the capability you want."

Inquiring lines that read this note 206

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.

What prevents language models from performing systematic logical reasoning? Does pretraining establish the ceiling for what reward learning can improve? Can smaller specialized models match frontier models on key metrics? When do simpler collaborative filtering approaches outperform complex LLM recommenders? Can minimal training unlock latent reasoning already present in base models? Can inference-time computation adaptively substitute for static model capacity? How do knowledge graph structures enable efficient multi-hop reasoning and retrieval? Does intelligent routing among smaller models outperform training larger models? When does parallel reasoning outperform sequential reasoning with the same token budget? What prediction granularity best trains models to generate reliable reasoning? How do thinking tokens exhibit diminishing returns in reasoning? Does chain-of-thought reasoning reveal how models actually think or merely imitate reasoning? What capabilities differentiate diffusion from autoregressive language models? How does fine-tuning trade off accuracy against reasoning quality? Can AI research automation sustain progress through accelerating feedback loops? Can confidence signals reliably detect flawed reasoning in language models? Can latent reasoning match or exceed explicit reasoning performance? How does model capacity affect learning performance on diverse downstream tasks? What are the fundamental limits of prompting for language models? Does scaling reasoning capability create fundamental tradeoffs in control and reliability? How effectively can test-time voting aggregate diverse reasoning samples? When do multi-agent systems improve over single frontier models? Can AI systems evade safety evaluations through reasoning manipulation? What limits language model accuracy in evaluating ideas? Can recurrent computation unlock reasoning capabilities that fixed-depth models cannot? How does policy entropy collapse limit scaling of reasoning-focused reinforcement learning? How should retrieval strategies adapt to multi-step reasoning demands? Why do training associations persist despite contradictory contextual information? How do neural networks learn compositional structure from training? Can reasoning models use reflection to correct their initial outputs? Does augmenting symbolic reasoning improve LLM logical reasoning ability? What makes reasoning traces effective supervision even when they're incorrect? How do reward signal properties affect model reasoning and safety? How do sequence length and task type interact with sparsity tolerance? How do real-world evaluations reveal AI capabilities that benchmarks hide? What explains the gap between benchmark scores and true reasoning capability? Which reinforcement learning modifications most improve dialogue quality in language models? Why does AI verification capability persistently exceed generation capability? What makes process supervision effective for training complex reasoning models? Can reasoning traces reveal actual model reasoning versus plausible output? Does reinforcement learning create genuinely new reasoning capabilities or only refine existing ones? How do training data quality and composition affect downstream model performance? How does diversity prevent model convergence on superficial patterns? How do curriculum design and feedback approaches affect model learning? Can AI systems perform peer review as effectively as humans?

Related concepts in this collection 7

This note in its neighbourhood — explore the map, then jump to a related concept in the list below.

Concept map
22 direct connections · 243 in 2-hop network ·dense cluster Open in graph ↗
Can non-reasoning models catch up with more comp… Can inference compute replace scaling up model siz… How do internal and external test-time scaling com… What makes test-time training actually work in pra… Can language models learn grammar from child-scale… Do base models already contain hidden reasoning ab… Does RLVR actually expand what models can reason a… Does RL post-training create reasoning or just dep…

Click a node to walk · click center to open · click Open in graph to see this note in the full knowledge graph

your link semantically near linked from elsewhere

Related papers in this collection 8

Papers most semantically related to this note, ranked by cosine similarity in the embedding space.

Search by related questions 4

Suggested questions this note speaks to — click to search the collection, or type your own.

Can targeted activation steering surface latent reasoning in base models?Why does imitation learning create a ceiling for reasoning capability?What makes reasoning-specific post-training different from standard parameter scaling?How much does training composition affect syntactic versus reasoning performance?
Original note title

non-reasoning models cannot match reasoning models even with unlimited inference budget