List your company
ResearchFormerly WizWand

Model Performance Prediction on DeepSeek Model Families (Hold-out)

0.02MAETailoredBench
-0.49382.974356.44259.91065Feb 8, 2026
Updated 8mo ago

Evaluation results

MethodLinks
TailoredBenchTarget Model=deepseek-llm-67b-base2026.020.02
SparseEvalTarget Model=deepseek-llm-67b-chat2026.020.156
SparseEvalTarget Model=deepseek-coder-6.7b-instruct2026.020.264
SparseEvalTarget Model=deepseek-llm-67b-base2026.020.73
SparseEvalTarget Model=deepseek-llm-7b-chat2026.020.738
TailoredBenchTarget Model=deepseek-moe-16b-base2026.020.926
TailoredBenchTarget Model=deepseek-coder-6.7b-instruct2026.021.096
SparseEvalTarget Model=deepseek-coder-1.3b-instruct2026.021.392
TailoredBenchTarget Model=deepseek-llm-67b-chat2026.021.448
SparseEvalTarget Model=deepseek-moe-16b-base2026.021.535
gp-IRTTarget Model=deepseek-llm-67b-chat2026.021.595
gp-IRTTarget Model=deepseek-coder-7b-instruct-v1.52026.021.681
SparseEvalTarget Model=deepseek-math-7b-base2026.021.777
gp-IRTTarget Model=deepseek-coder-1.3b-instruct2026.021.873
gp-IRTTarget Model=deepseek-llm-67b-base2026.021.882
gp-IRTTarget Model=deepseek-math-7b-instruct2026.021.95
gp-IRTTarget Model=deepseek-math-7b-base2026.022.019
gp-IRTTarget Model=deepseek-coder-6.7b-base2026.022.083
SparseEvalAggregate=Average MAE2026.022.09
SparseEvalTarget Model=deepseek-coder-6.7b-base2026.022.132
gp-IRTTarget Model=deepseek-moe-16b-base2026.022.244
gp-IRTTarget Model=deepseek-coder-6.7b-instruct2026.022.304
gp-IRTTarget Model=deepseek-math-7b-rl2026.022.374
TailoredBenchTarget Model=deepseek-math-7b-base2026.022.459
SparseEvalTarget Model=deepseek-math-7b-rl2026.022.542
gp-IRTTarget Model=deepseek-llm-7b-chat2026.022.624
gp-IRTAggregate=Average MAE2026.022.657
TailoredBenchTarget Model=deepseek-llm-7b-chat2026.022.816
TailoredBenchTarget Model=deepseek-math-7b-rl2026.023.407
SparseEvalTarget Model=deepseek-coder-7b-instruct-v1.52026.023.7
TailoredBenchTarget Model=deepseek-math-7b-instruct2026.023.712
SparseEvalTarget Model=deepseek-moe-16b-chat2026.023.74
TailoredBenchTarget Model=deepseek-coder-6.7b-base2026.023.751
SparseEvalTarget Model=deepseek-math-7b-instruct2026.023.81
TailoredBenchTarget Model=deepseek-coder-7b-instruct-v1.52026.024.182
TailoredBenchAggregate=Average MAE2026.024.216
SparseEvalTarget Model=deepseek-llm-7b-base2026.024.616
SparseEvalAggregate=Max MAE2026.024.616
gp-IRTTarget Model=deepseek-llm-7b-base2026.025.71
gp-IRTTarget Model=deepseek-moe-16b-chat2026.026.195
gp-IRTAggregate=Max MAE2026.026.195
TailoredBenchTarget Model=deepseek-coder-1.3b-instruct2026.026.464
TailoredBenchTarget Model=deepseek-llm-7b-base2026.0211.662
TailoredBenchTarget Model=deepseek-moe-16b-chat2026.0212.865
TailoredBenchAggregate=Max MAE2026.0212.865