List your company

Outcome correctness on TPS-CalcBench core set v2

51.9Exact Matchgpt-5.2
-2.07611.93725.9539.963Apr 20, 2026

Evaluation results

MethodLinks
gpt-5.22026.0451.9299.89.387.695
gpt-5.12026.0448.127.9101485.295.7
glm-52026.0447.129.313.89.886.2100
deepseek-v3.12026.0443.128.11018.881.787.6
aws.claude-opus-4.52026.0441.924.314.519.385100
Doubao-Seed-1.82026.0438.628.114.518.875.7100
glm-4.72026.0437.624.313.624.576.4100
aws.claude-sonnet-4.52026.0433.628.314.823.372.1100
aws.claude-haiku-4.52026.0424.328.118.828.866.2100
gemini-3-flash-preview2026.049.813.824.352.152.6100
MiniMax-M2.52026.040514.880.224.397.1
MiniMax-M2.12026.0404.59.885.718.897.9