invokeable()
← comparisons / claude-opus-4-8-vs-claude-opus-5

Claude Opus 4.8 vs Claude Opus 5

Side-by-side from the Invokeable directory — provider-attributed pricing, canonical IDs, draft data flagged. Open as a shareable card →

01Head to head

Claude Opus 4.8Claude Opus 5
CreatorAnthropicAnthropic
Released2026-05-282026-07-24
Context1M1M
Max output128K128K
Input $/M——
Output $/M——
Knowledge cutoff2026-012026-05
Modalitiestext + image → texttext + image → text
Capabilitiestool-use, extended-thinkingtool-use, extended-thinking
ProvidersAnthropic APIAnthropic API

02Benchmarks · vendor-reported unless verified

Claude Opus 4.8Claude Opus 5
DeepSWE—68.8%
FrontierCode (mean@5)—53.4
Humanity's Last Exam (accuracy)49.8—
OSWorld-Verified (success rate)83.4—
SWE-bench Multilingual—89.5%
SWE-bench Multimodal—59.4%
SWE-bench Pro69.2%79.2%
SWE-bench Verified (resolved)88.696
Terminal-Bench (success rate)74.6—