List your company

General Capability on 8 capability benchmarks Aggregate

67.14Average CapabilityOriginal Weight Model
46.527251.878657.2362.5814Apr 2, 2025May 20, 2025Jul 8, 2025Aug 26, 2025Oct 14, 2025Dec 2, 2025Jan 20, 2026

Evaluation results

MethodLinks
Original Weight ModelModel=Llama3 8B Instruct2025.0467.1473.9
NPOModel=Llama3 8B Instruct2025.0466.7171.65
CBModel=Llama3 8B Instruct2025.0466.5881.61
RAM+Model Category=Merged Models2026.0166.55—
CBModel=Llama3 8B Instruct, safety-tuned=true2025.0466.4775.87
REPBENDModel=Llama3 8B Instruct2025.0465.983.14
RAMModel Category=Merged Models2026.0164.82—
DARE+TAModel Category=Merged Models2026.0163.33—
DARE+TIESModel Category=Merged Models2026.0162.47—
TIESModel Category=Merged Models2026.0160.02—
Original Weight ModelModel=Mistral 7B Instruct v0.22025.0459.1863.81
CBModel=Mistral 7B Instruct v0.22025.0458.9777.34
CBModel=Mistral 7B Instruct v0.2, safety-tuned=true2025.0458.9373.62
TAModel Category=Merged Models2026.0158.28—
MemAgent (Memory)Model Category=Base and Task Models2026.0157.77—
REPBENDModel=Mistral 7B Instruct v0.22025.0457.6881.23
TAModel=Llama3 8B Instruct2025.0457.4370.71
RMUModel=Llama3 8B Instruct2025.0454.8472.9
ToolRL (Tool)Model Category=Base and Task Models2026.0154.16—
NPOModel=Mistral 7B Instruct v0.22025.0453.9474.52
TAModel=Mistral 7B Instruct v0.22025.0453.172.96
FisherModel Category=Merged Models2026.0152.2—
CURE (Coding)Model Category=Base and Task Models2026.0149.35—
BaseModel Category=Base and Task Models2026.0148.7—
R2D2Model=Mistral 7B Instruct v0.2, safety-tuned=true2025.0448.4472.67
RMUModel=Mistral 7B Instruct v0.22025.0447.3271.85