← Claude Opus 4.8 · API · High effort (default)
Evidenceacceptedvendor_reported
Humanity's Last Exam
Claude Opus 4.8 System Card — HLE with tools
ev_2026_07_opus48_hle_vendor
Result
57.9%
Confidence B
- Tested
- 2026-05-28
- Runs
- 1
Configuration
How it was measured
- Exact model
- claude-opus-4-8
- Reasoning
- high
- Agent harness
- none_or_vendor_hle_tools
- Tool access
- tools enabled
- Prompting
- vendor HLE with tools
SECONDARY vendor tool-enabled HLE.
Caveats
Known limitations
- Tool-enabled; not closed-book.
Weighting
Evidence weight factors
The engine weights each evidence row by these factors (0–1) when fusing scores — higher is more trusted.
relevance0.90
quality0.55
recency0.95
health0.90
comparability0.80
reliability0.65
Benchmark
Humanity's Last Exam
Frontier closed-ended academic benchmark across many subjects; often reported with and without tools.
Benchmark source →