Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters

Why this source is here

The inference-time compute trade-off against parameter scaling.

Verification

Content-verified on 2026-07-27: the canonical source and its title were resolved during the Atlas review. This is not an endorsement of the source’s argument.

Authors

Snell et al.

Identifier

arXiv:2408.03314