mlsanityData to 7 Oct 2026

Temperature-Dependent Performance of Prompting Strategies in Extended Reasoning Large Language Models

arXiv 2604.08563 ↗First posted 18 Mar 2026cs.CL cs.AI cs.LGcited by 0

On the map: LLM reasoning → Prompting and chain-of-thought reasoning

Abstract (arXiv)

Extended reasoning models represent a transformative shift in Large Language Model (LLM) capabilities by enabling explicit test-time computation for complex problem solving. However, the optimal configuration of sampling temperature and prompting strategy for these systems remains largely underexplored. We systematically evaluate chain-of-thought and zero-shot prompting across four temperature settings (0.0, 0.4, 0.7, and 1.0) using Grok-4.1 with extended reasoning on 39 mathematical problems from AMO-Bench, a challenging International Mathematical Olympiad-level benchmark. We find that zero-shot prompting achieves peak performance at moderate temperatures, reaching 59% accuracy at T=0.4 and T=0.7, while chain-of-thought prompting performs best at the temperature extremes. Most notably, the benefit of extended reasoning increases from 6x at T=0.0 to 14.3x at T=1.0. These results suggest that temperature should be optimized jointly with prompting strategy, challenging the common practice of using T=0 for reasoning tasks.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning. Builds on 17 of its 19 references that are in the map. First posted on arXiv 18 Mar 2026.

Builds on

17 of 19 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025Prompting Science Report 2: The Decreasing Value of Chain of Thought in Prompting8 Jun
  2. 2022Large Language Models are Zero-Shot Reasoners24 May
  3. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Models28 Jan

Other references · 14

  1. 2023GPT-4 Technical Reportcited by 18,132
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  4. 2019The Curious Case of Neural Text Degenerationcited by 2,881
  5. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 895
  6. 2024The Effect of Sampling Temperature on Problem Solving in Large Language Modelscited by 168
  7. 2025A Survey of Frontiers in LLM Reasoning: Inference Scaling, Learning to Reason, and Agentic Systemscited by 108
  8. 2024LLM as a Mastermind: A Survey of Strategic Reasoning with Large Language Modelscited by 97
  9. 2025Sampling-Efficient Test-Time Scaling: Self-Estimating the Best-of-N Sampling in Early Decodingcited by 62
  10. 2025AMO-Bench: Large Language Models Still Struggle in High School Math Competitionscited by 36
  11. 2025Optimizing Temperature for Language Models with Multi-Sample Inferencecited by 33
  12. 2023Chain-of-Thought in Neural Code Generation: From and For Lightweight Language Modelscited by 27
2 more
  1. 2025On the Role of Temperature Sampling in Test-Time Scalingcited by 10
  2. 2024Automatic Generation of Model and Data Cards: A Step Towards Responsible AIcited by 9

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024T2 of Thoughts: Temperature Tree Elicits Reasoning in Large Language Models
  2. 2025Reasoning Capabilities of Large Language Models on Dynamic Tasks
  3. 2025Reasoning Strategies in Large Language Models: Can They Follow, Prefer, and Optimize?
  4. 2026Diverge to Induce Prompting: Multi-Rationale Induction for Zero-Shot Reasoning
  5. 2026Fragile Thoughts: How Large Language Models Handle Chain-of-Thought Perturbations
  6. 2025Exploring LLM Reasoning Through Controlled Prompt Variations
  7. 2024Towards Generalist Prompting for Large Language Models by Mental Models
  8. 2024Large Language Models are Contrastive Reasoners