Updated minutes ago
Llama 3.2 1B
Meta · dense · 1.24B parameters · 131,072 context
Quality38.0
Architecture Details
TypeDENSE
Total Parameters1.24B
Active Parameters1.24B
Layers16
Hidden Dimension2,048
Attention Heads32
KV Heads8
Head Dimension64
Vocab Size128,256
Memory Requirements
BF16 Weights
2.5 GB
FP8 Weights
1.2 GB
INT4 Weights
0.6 GB
KV-Cache per Token32768 bytes
Activation Estimate0.30 GB
Fits on (single-node)
B200 SXM BF16B100 SXM BF16GB200 NVL72 (per GPU) BF16GB300 NVL72 (per GPU) BF16H200 SXM BF16H100 SXM BF16H100 PCIe BF16H100 NVL BF16
GPU Recommendations
RTX 4070 Tioptimal
BF16 · 1 GPU · vllm
90/100
score
Throughput
1.0K tok/s
Cost/Month
$237
Cost/M Tokens
$0.09
RTX 3080optimal
BF16 · 1 GPU · vllm
90/100
score
Throughput
1.6K tok/s
Cost/Month
$133
Cost/M Tokens
$0.03
RTX 4060optimal
BF16 · 1 GPU · vllm
90/100
score
Throughput
562.6 tok/s
Cost/Month
$209
Cost/M Tokens
$0.14
API Pricing Comparison
| Provider | Input $/M | Output $/M | Badges |
|---|---|---|---|
| together | $0.03 | $0.03 | Cheapest |
| fireworks | $0.10 | $0.10 |
Quality Benchmarks
MMLU49.3
HumanEval22.0
GSM8K44.4
MT-Bench62.0
Capabilities
Features
✓ Tool Use✗ Vision✓ Code✓ Math✗ Reasoning✓ Multilingual✓ Structured Output
Supported Frameworks
vllmsglangtgitensorrt-llmollama
Supported Precisions
BF16 (default)FP8INT4