Atlas

Models

← All models

Agnes 2.5 Pro Beta

Sapiens AI

Sapiens AI's proprietary Agnes 2.5 Pro Beta release, evaluated separately from Pro Alpha by Artificial Analysis, which reports text and image inputs. Its API documents an optional thinking request field for enabling Thinking mode. Public evidence does not establish its architecture.

Benchmark scores

BenchmarkScore
AA-LCR v1.183.0
AA-Omniscience Index-10.5
Agents' Last Exam (ALE)21.7
Agents' Last Exam ALE-CLI Pass Rate25.7
Agents' Last Exam ALE-CLI Score50.2
Agents' Last Exam Full-Spectrum Pass Rate23.6
Agents' Last Exam Full-Spectrum Score41.5
Agents' Last Exam Last-Exam Pass Rate5.3
Agents' Last Exam Last-Exam Score12.6
Agents' Last Exam Near-Term Pass Rate31.3
Agents' Last Exam Near-Term Score60.2
Agents' Last Exam Overall Score42.7
Agents' Last Exam Unlicensed Full-Spectrum Pass Rate24.0
Agents' Last Exam Unlicensed Full-Spectrum Score43.7
Agents' Last Exam Unlicensed Last-Exam Pass Rate6.7
Agents' Last Exam Unlicensed Last-Exam Score15.9
Agents' Last Exam Unlicensed Near-Term Pass Rate31.3
Agents' Last Exam Unlicensed Near-Term Score60.2
Agents' Last Exam Unlicensed Overall Pass Rate23.4
Agents' Last Exam Unlicensed Overall Score46.1
Artificial Analysis Omniscience Accuracy16.8
Artificial Analysis Omniscience Hallucination Rate32.8
Artificial Analysis Progress Index v1299
Capability157.8
CritPt (avg@5)15.7
GDPval-AA v21376
GDPval-AA v2.11336
GPQA Diamond (Artificial Analysis avg@5)90.5
Humanity's Last Exam (Text-Only)37.5
MMMU Pro75.4
Loading Atlas data…