Private AI
An open-weight 21B parameter model released under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI's Harmony response format and supports reasoning level configuration, fine-tuning, and agentic capabilities including function calling, tool use, and structured outputs.
Added Aug 5, 2025
Model weightsContext Window
128.0K
Max Output
16.4K
Input Price (Auto)
$0.032/1M
Output Price (Auto)
$0.14/1M
Cache Read (Auto)
$0.032/1M
Capabilities
Performance metrics and benchmarks
Sourced from Artificial Analysis.
Intelligence Index
14.9
Choose explicit providers for this model. Auto routing remains available as the default option.
Loading provider options…
Coding Index
20.7
GPQA Diamond
Graduate-level scientific reasoning
68.8%
Better than 50% of models compared
HLE
Humanity's Last Exam
9.8%
Better than 60% of models compared
IFBench
Instruction-following benchmark
65.1%
Better than 77% of models compared
T²-Bench Telecom
Conversational AI agents in dual-control scenarios
60.2%
Better than 57% of models compared
AA-LCR
Long context reasoning evaluation
30.7%
Better than 41% of models compared
SciCode
Python programming for scientific computing
34.4%
Better than 52% of models compared
Terminal-Bench Hard
Agentic coding and terminal use
10.6%
AIME 2025
American Invitational Mathematics Examination 2025
89.3%
Better than 88% of models compared
MMLU-Pro
Professional and academic subject knowledge
74.8%
Better than 47% of models compared
Last updated Aug 4, 2026
Artificial AnalysisBetter than 45% of models compared
LiveCodeBench
Contamination-free coding benchmark
77.7%
Better than 89% of models compared
