BenchAtlas

Rankings / Fireworks AI

Deepseek V3p2

release date unknown

BenchAtlas Index

as of 2026-07-14
Not enough benchmark coverage for a composite score (needs ≥4 families across ≥3 categories). The per-benchmark evidence below stands on its own — sparse coverage is an evidence gap, not a low score.

Benchmark evidence

14 results
knowledge science
GPQA Diamond
implementation=vals-ai
high effort80.3 %independentVals AI
GPQA Diamond
implementation=vals-ai
no reasoning76.3 %independentVals AI
MMLU-Pro
implementation=vals-ai
high effort84.9 %independentVals AI
MMLU-Pro
implementation=vals-ai
no reasoning83.1 %independentVals AI
professional
CaseLawhigh effort55.4 %independentVals AI
CorpFinhigh effort51.0 %independentVals AI
CorpFinno reasoning47.9 %independentVals AI
LegalBenchhigh effort76.1 %independentVals AI
LegalBenchno reasoning66.0 %independentVals AI
MedQAhigh effort93.9 %independentVals AI
MedQAno reasoning89.5 %independentVals AI
TaxEvalhigh effort68.2 %independentVals AI
reasoning math
AIME
implementation=vals-ai
high effort84.6 %independentVals AI
AIME
implementation=vals-ai
no reasoning64.8 %independentVals AI