VECTOR WIREAI INTELLIGENCE
UTC
Refresh Models Deals Regulatory Sources

Claude Opus 4.8 vs Step 5 Preview

11 SHARED BENCHMARKS

Across 11 shared benchmarks, Claude Opus 4.8 scores higher on 6 and Step 5 Preview on 5. The widest gap is AA-Omniscience, where Claude Opus 4.8 scores 28.8 against 16.4. Step 5 Preview is the cheaper of the two on tracked API pricing ($1.00 against $5.00 per million input tokens).

ANTHROPICVSSTEPFUN11 SHARED65 HEAD-TO-HEADNEWEST SCORE ADDED
AA-LCR77.788.3
critpt20.920.9
gdpval49.553.3
HLE57.946.5
MMMU-Pro78.976.4
scicode54.458.9
terminalbenchV4021.733.3

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (anthropic-official, direct), otherwise the lowest tracked offer.

Compare Claude Opus 4.8 withALL PAIRINGS →