Refresh Models —Deals —Regulatory —Sources —

Claude Haiku 4.5 vs HyperNova 60B 2605 (high, based on gpt-oss-120b)

17 SHARED BENCHMARKS

Across 17 shared benchmarks, Claude Haiku 4.5 scores higher on 12 and HyperNova 60B 2605 (high, based on gpt-oss-120b) on 5. The widest gap is AA Agentic Index, where Claude Haiku 4.5 scores 32.6 against 6.8.

ANTHROPICVSOPENAI17 SHARED12–5 HEAD-TO-HEAD

BenchmarkClaude Haiku 4.5MARGINHyperNova 60B 2605 (high, based on gpt-oss-120b)

AA Agentic Index32.66.8

AA Intelligence29.618.3

AA-LCR70.335.3

AA-Omniscience-4.2-54.6

Artificial Analysis Coding Index43.923.2

GPQA Diamond7373.3

HLE9.716.6

IFBench54.366.5

OmniScience Accuracy17.415.2

OmniScience Non-Hallucination75.317.7

TauBench V3 - Banking9.15.6

Terminal-Bench 2.144.218.4

Terminal-Bench Hard27.323.5

τ²-Bench Telecom (AA run)54.763.2

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.