Agnes 3.0 Flash
Parameters
33B
Context Length
262K
Modality
Text
Architecture
Dense
KV Cache Rate
288 KB/tok
License
Apache 2.0
Release Date
11 Sept 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Input: $0.05 · Output: $0.15
Hardware Sizing
Precision
Context
Workload Memory: 20.12 GB
INT4 · 8K context
Base Weights: 15.37 GB
KV Cache: 2.25 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
1x NVIDIA L4 24GB
20.1 / 21.6 GB usable
2x RTX 4090 24GB
20.1 / 43.2 GB usable · TP 2
Apple M3 Pro 36GB
20.1 / 27.0 GB usable
Determine requirements for your specific workload:
Open in VRAM Calculator
Architecture Diagram
Evaluation Benchmarks
No evaluation benchmarks for Agnes 3.0 Flash available.
Rankings
Overall Rank
-
Coding Rank
-
About Agnes 3.0 Flash
Agnes 3.0 Flash is a lightweight conditional generation model designed for fast multimodal and conversational tasks. It offers optimized throughput for latency-critical deployment environments.
Technical Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
24
Key-Value Heads
4
Attention Head Dimension
256
Position Embedding
ROPE
RoPE Theta
10,000,000
Linear Attention
Yes
Linear Attention Ratio
75.0%
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Hidden Dimension Size
5,120
Number of Layers
72
FFN Intermediate Size (Dense)
17,408
Multi-Token Prediction Heads
1
Tokenizer
Vocabulary Size
248,320
Resources
About Agnes 3.0
The Agnes 3.0 model family developed by Agnes AI.
Other Agnes 3.0 Models
- No related models available