Classroom Final Exam: An Instructor-Tested Reasoning Benchmark
On the map: LLM reasoning → Mathematical reasoning
Abstract (arXiv)
We introduce CFE-Bench (Classroom Final Exam), a multimodal benchmark for evaluating the reasoning capabilities of large language models across more than 20 STEM domains. CFE-Bench is curated from repeatedly used, authentic university homework and exam problems, paired with reference solutions provided by course instructors. CFE-Bench remains challenging for frontier models: the newly released Gemini-3.1-pro-preview achieves 59.69% overall accuracy, while the second-best model, Gemini-3-flash-preview, reaches 55.46%, leaving substantial room for improvement. Beyond aggregate scores, we conduct a diagnostic analysis by decomposing instructor reference solutions into structured reasoning flows. We find that while frontier models often answer intermediate sub-questions correctly, they struggle to reliably derive and maintain correct intermediate states throughout multi-step solutions. We further observe that model-generated solutions typically contain more reasoning steps than instructor solutions, indicating lower step efficiency and a higher risk of error accumulation. Data and code are available at https://github.com/Analogy-AI/CFE_Bench.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Mathematical reasoning, one of 17 directions in LLM reasoning. Builds on 21 of its 43 references that are in the map. First posted on arXiv 23 Feb 2026.
Builds on
21 of 43 references.
Closest ancestors · 1
Same direction, newest first
Other references · 20
- 2025Qwen3 Technical Reportcited by 8,195
- 2020Measuring Massive Multitask Language Understandingcited by 7,768
- 2023Let's Verify Step by Stepcited by 3,598
- 2025Gemma 3 Technical Reportcited by 1,686
- 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiencycited by 1,419
- 2025gpt-oss-120b & gpt-oss-20b Model Cardcited by 1,258
- 2023Gemini: A Family of Highly Capable Multimodal Modelscited by 810
- 2025DeepSeek-V3.2: Pushing the Frontier of Open Large Language Modelscited by 755
- 2025Kimi K2: Open Agentic Intelligencecited by 523
- 2026GLM-5: from Vibe Coding to Agentic Engineeringcited by 502
- 2025Humanity's Last Examcited by 467
- 2026Kimi K2.5: Visual Agentic Intelligencecited by 449
8 more
- 2024Measuring short-form factuality in large language modelscited by 345
- 2026Ministral 3cited by 88
- 2025Winning Gold at IMO 2025 with a Model-Agnostic Verification-and-Refinement Pipelinecited by 79
- 2025No Free Labels: Limitations of LLM-as-a-Judge Without Human Groundingcited by 50
- 2026FrontierScience: Evaluating AI's Ability to Perform Expert-Level Scientific Taskscited by 42
- 2024How predictable is language model benchmark performance?cited by 36
- 2025HiPhO: How Far Are (M)LLMs from Humans in the Latest High School Physics Olympiad Benchmark?cited by 18
- 2025Beyond Consensus: Mitigating the Agreeableness Bias in LLM Judge Evaluationscited by 16
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2024CLR-Bench: Evaluating Large Language Models in College-level Reasoning
- 2025R-Bench: Graduate-level Multi-disciplinary Benchmarks for LLM & MLLM Complex Reasoning Evaluation
- 2025The Quest for Efficient Reasoning: A Data-Centric Benchmark to CoT Distillation
- 2023Have LLMs Advanced Enough? A Challenging Problem Solving Benchmark For Large Language Models
- 2025UGMathBench: A Diverse and Dynamic Benchmark for Undergraduate-Level Mathematical Reasoning with Large Language Models
- 2024Functional Benchmarks for Robust Evaluation of Reasoning Performance, and the Reasoning Gap
- 2025The Ouroboros of Benchmarking: Reasoning Evaluation in an Era of Saturation
- 2025EffiReason-Bench: A Unified Benchmark for Evaluating and Advancing Efficient Reasoning in Large Language Models