MGSM (Multilingual Grade School Math (MGSM) benchmark)

Explore score history ↗

MGSM (Multilingual Grade School Math (MGSM) benchmark) is an AI benchmark in the Benchmark Radar catalog.

Released: 2022-10-07
Openness: unknown
Importer: Benchmark Saturation

Source provenance

Original evidence

MGSM (Multilingual Grade School Math (MGSM) benchmark) results and reported scores

Reported benchmark scores · Benchmark Saturation

Reported scores: 5

Model	Organization	Reported value	Reported	Evidence
Claude Opus 4	—	93.8	—	evidence
Claude Opus 4.1	—	94.4	—	evidence
Claude Sonnet 4	—	93.0	—	evidence
Claude Sonnet 4.5	—	94.3	—	evidence
o4 Mini	—	93.4	—	evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.

Which sources cite MGSM (Multilingual Grade School Math (MGSM) benchmark)?