elephant-1 · IQ1_S
This page aggregates 1 real-world runs of elephant-1 (IQ1_S) on NVIDIA RTX 3090 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.
163.87 tok/s
Decode
Decode speed
748 tok/s
Prefill
Prefill speed
1.44 s
TTFT
Time to first token
10.04 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
Other
Source platforms
13 days ago
Last verified
Performance
- Ollama · Q4_K_S Decode 69.84 · Prefill 337.36 ·
- Strata · IQ1_S (current)Decode 163.87 · Prefill 748 ·
Core figures
- Decode (avg)
- 163.87 tok/s
- Prefill (avg)
- 748 tok/s
- TTFT (avg)
- 1.44 s
- VRAM (avg)
- 10.04 GB
- MTP acceptance rate
- 0.66
- TTFB
- 1.31 GB
- Power draw
- 307.2 W
Configuration
Member-level fields are taken from the most recent run
- Model
- elephant-1
- Quantization
- IQ1_S
- Framework
- Strata
- Version
- v0.26
- Context length
- 4096 tokens
- Batch size
- 512
- GPU layers
- 8
- Flash Attention
- Off
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX 3090
- Nominal VRAM
- 24 GB
- Measured VRAM (avg)
- 10.04 GB
- OS
- Ubuntu 24.04
- Driver
- 572.83
- CUDA
- 12.8
- Power draw
- 307.2 W
Get started
Original model resources
Base model resources may not match this quantization. Use the verified deployment weights above when available.
Sources & evidence
1 measured records in total, each traceable to its original source
v0.26 · Ubuntu 24.04 · CUDA 12.8 · 4096 ctx
163.87 tok/s
Decode
748 tok/s
Prefill
1.44 s
TTFT
10.04 GB
VRAM
0.66
MTP
307.2 W
Power draw
L2 batch seed 2026-10-22: Strata on hw_id=2 running model_id=95 at IQ1_S