LinAlg-Bench - 5x5 Matrices: leaderboard

Exact linear algebra on 220 SymPy-certified integer 5x5 matrix problems from LinAlg-Bench (2026): trace, transpose, products, matrix powers, nullity, rank, determinants and eigenvalues, solved zero-shot with shown work and no tools. Accuracy in percent; higher is better.

Metric: Overall accuracy (%) on the 220 integer-entry 5x5 matrix problems (nine task types: trace, transpose, matrix-vector product, multiplication, matrix power, nullity and rank, 20 problems each, plus 50 determinants and 30 eigenvalue problems), zero-shot chain-of-thought prompt at temperature 0 with tools and code execution disabled, one run per problem, answers checked against SymPy ground truth; the table prints the share rounded to whole points; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1O187#403
2Gemini 3.1 Pro (Preview)86#99
3DeepSeek V3.282#419
4Qwen 3 235B A22B 2507 Instruct82#616
5GPT-5.275#185
6Mistral Large 367#832
7Claude Sonnet 4.562#261
8GPT-4o57#751
9Llama 3.3 70B Instruct53#1385
10Qwen 2.5 72B Instruct49#1117

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=linalg-bench-5x5-matrices · How It Works · Data refreshed daily, snapshot 2026-11-21.