Line of inquiry
Inquiring lines›How do training data and architect…›How should model design balance ar…›this line of inquiry
How does scaling test-time compute compare to increasing model size?
A broader line of inquiry — a family of 59 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 59
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- How does test-time compute substitute for model parameter scaling?
- Does test-time compute actually substitute for having larger model parameters?
- Can test-time compute scaling substitute for larger model parameters?
- Where does inference compute stop substituting for model capacity?
- Do models excel at reasoning depth or memory breadth when scaling test time compute?
- Does inference-time compute improve pretraining data efficiency in practice?
- What patterns emerge across test-time scaling and reasoning architectures?
- Can test-time compute on smaller models replace larger model inference?
- Can reasoning models outperform non-reasoning models with more inference compute?
- How do internal versus external test-time scaling approaches differ from precomputation strategies?
- Does trading model size for inference steps improve overall efficiency scaling?
- Can test-time scaling prioritize genuine reasoning over pattern matching?
- Does more inference compute help reasoning models match specialized domain performance?
- Does inference-time compute scaling require explicit reasoning traces or verifiable rewards?
- How do architectural choices reduce inference cost independent of training compute?
- Does more inference compute help close gaps between different training regimes?
- Does decoupling reasoning reduce inference cost more than sequential scaling?
- Does test-time compute scaling work for agentic deep research tasks?
- How does inference compute substitution affect the training parameter scaling trade-off?
- Does flexible inference-time compute scaling through looping improve efficiency further?
- Can post-thinking compute on memory reduce query-time reasoning costs?
- Can test-time compute fully replace scaling model parameters on hard problems?
- Can test-time scaling work through retrieval rather than reasoning?
- Why does architecture matter more than training compute for inference efficiency?
- What efficiency gains could prevent an inference compute bottleneck?
- Can memory and test-time compute scale together as a single axis?
- Can weaker models match stronger ones with sufficient search and reasoning budget?
- How much inference efficiency do we gain by eliminating self-correction passes?
- Where does sleep-time compute fit in the taxonomy of test-time scaling?
- When does the right constraint beat additional model capacity?
- How do sequential and parallel compute primitives differ in test-time scaling?
- How does test-time scaling relate to token budget in agentic deep research?
- How do architectural changes affect inference efficiency independent of model size?
- How does distributed inference enable models to run on weak hardware?
- Can sleep-time compute reduce latency demands during model inference?
- Does inference demand for memory-hungry accelerators explain the shift toward memory-dominant bills of materials?
- What inference-time scaling benefits emerge from reasoning before each prediction?
- Are inference speed and context length also within parity on consumer GPU hardware?
- What architectural variables most improve inference efficiency today?
- How do sleep-time and post-completion methods reduce inference latency?
- What is the trade-off between parallel and sequential scaling at test time?
- What happens when models overthink during test-time search?
- How much inference compute does panel-of-judges evaluation actually cost?
- What deployment tradeoffs emerge between single-pass and multi-pass inference adaptation?
- What is the latency and compute cost of running memory inference?
- How does task structure determine optimal test-time compute allocation?
- How does precomputing context reasoning reduce latency in stateful applications?
- How does the inference steps dial compare to test-time compute trade-offs in language models?
- Does test-time compute substitute for parameter scaling at stable rates?
- Can offline context optimization reduce test-time latency like sleep-time compute?
- How does conditional memory compare to conditional computation for efficiency?
- Can test-time scaling compound through memory consolidation into a new scaling law?
- How does spending offline compute affect wake-time prediction latency?
- How do conditional scaling laws incorporate hardware into architecture choices?
- How does OpenAI distinguish between training compute and research compute internally?
- What are the computational trade-offs between training-time vs inference-time consistency correction?
- How do KV cache pruning and subproblem contraction both free reasoning capacity?
- How should we measure and report serial compute separately?
- How does the three-component definition apply to test-time scaling laws?