Nous ResearchOpen weightsAuto-detected

Hermes 4 70B

Name: Hermes 4 70B benchmark results
Creator: BenchLeader

Best configuration ranks #486 of 610 on the BenchLeader Index at 41.4 ±3.4 (thinking reasoning effort).

Blended price: –
Output speed: –
First answer: –
Context: 128k

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index41
Agents & tools38
Knowledge41
Instruction following35
Long context30
Composite39

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.

Effort	Index	Rank	Speed	First answer	Chat reply cost	Categories
thinkingbest	41.4	#486	–	–	–	Agents & tools 38 · Composite 39 · Instruction following 35 · Knowledge 41 · Long context 30
default	40.2	#514	–	–	–	Agents & tools 34 · Composite 37 · Instruction following 33 · Knowledge 41 · Long context 27

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmark	thinking	default	Source
GPQA Diamond (AA)not in index	69.9%#287	49.1%#410	Artificial Analysis
Humanity's Last Exam (AA)not in index	8.8%#275	3.6%#499	Artificial Analysis

Agents & tools

Benchmark	thinking	default	Source
Terminal-Bench Hard	4.5%#270	0.0%#345	Artificial Analysis
τ²-Bench Telecom (AA)not in index	22.5%#297	21.6%#305	Artificial Analysis

Knowledge

Benchmark	thinking	default	Source
AA-Omniscience	-49.4#312	-48.6#305	Artificial Analysis

Instruction following

Benchmark	thinking	default	Source
IFBench	31.3%#338	29.0%#351	Artificial Analysis

Long context

Benchmark	thinking	default	Source
AA-LCR	9.7%#396	4.7%#415	Artificial Analysis

Composite

Benchmark	thinking	default	Source
AA Intelligence Index	7.9#372	6.7#434	Artificial Analysis