DeepSeek V3 vs DeepSeek V3-0324
Architecture Comparison
SpecDeepSeek V3DeepSeek V3-0324
TypeMOEMOE
Total Parameters671B685B
Active Parameters37B37B
Layers6161
Hidden Dimension7,1687,168
Attention Heads128128
KV Heads11
Context Length131,072131,072
Precision (default)BF16FP8
Total Experts256256
Active Experts88
Memory Requirements
PrecisionDeepSeek V3DeepSeek V3-0324
BF16 Weights1342.0 GB1370.0 GB
FP8 Weights671.0 GB685.0 GB
INT4 Weights335.5 GB342.5 GB
KV-Cache / Token31232 B31232 B
Activation Estimate3.00 GB3.00 GB
Minimum GPUs Needed (BF16)
H100 SXMN/AN/A
L40SN/AN/A
Quality Benchmarks
BenchmarkDeepSeek V3DeepSeek V3-0324
Overall8650
MMLU87.1N/A
HumanEval65.0N/A
GSM8K89.3N/A
MT-Bench87.0N/A
DeepSeek V3
MMLU
87.1
HumanEval
65.0
GSM8K
89.3
MT-Bench
87.0
DeepSeek V3-0324
Capabilities
FeatureDeepSeek V3DeepSeek V3-0324
Tool Use✓ Yes✓ Yes
Vision✗ No✗ No
Code✓ Yes✓ Yes
Math✓ Yes✓ Yes
Reasoning✗ No✓ Yes
Multilingual✓ Yes✓ Yes
Structured Output✓ Yes✓ Yes
API Pricing Comparison
Cheapest Output (DeepSeek V3)
$0.42/M
Input: $0.28/M
Cheapest Output (DeepSeek V3-0324)
$0.42/M
Input: $0.28/M
| Provider | DeepSeek V3 In $/M | Out $/M | DeepSeek V3-0324 In $/M | Out $/M |
|---|---|---|---|---|
| deepseek | $0.28 | $0.42 | $0.28 | $0.42 |
| together | $0.50 | $2.80 | $0.50 | $2.80 |
Recommendation Summary
- ‣DeepSeek V3 scores higher on overall quality (86 vs 50).
- ‣DeepSeek V3 has a smaller memory footprint (1342.0 GB vs 1370.0 GB BF16), making it easier to deploy on fewer GPUs.