Benchmark Almanac

Moonlight-16B-A3B

Moonlight-16B-A3B.

Moonshot AI Released 22 Feb 2025 Family Moonlight The lab's model card
11
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Moonshot AI Feb 2025

Moonlight-16B-A3B

Open weights Small

No Epoch Capabilities Index score yet
  • MMLU-Pro 42.4% #146 /151
  • MMLU 70% #96 /127
  • C-Eval 77.2% #32 /38

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
BIG-Bench Hard 65.2% #52 of 71 Model card
MMLU 70% #96 of 127 Model card
TriviaQA 66.3% #31 of 39 Model card
CMMLU 78.2% #14 of 17 Model card
C-Eval 77.2% #32 of 38 Model card
MMLU-Pro 42.4% #146 of 151 Model card

Math and science

BenchmarkScoreRankSource
GSM8K 77.4% #64 of 87 Model card
MATH 45.3% #69 of 93 Model card
CMATH 81.1% #5 of 6 Model card

Coding

BenchmarkScoreRankSource
MBPP 63.8% #47 of 66 Model card
HumanEval 48.1% #80 of 95 Model card

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.