DeepSeek V4 Pro 1.6T
8K/1K
B200
vLLM · FP4
MI355X
—
cannot reach @150
B300
$0.64250% cheaper than B200
Dynamo SGLang · FP4
GB200 NVL72
$1.13311% cheaper than B200
Dynamo SGLang · FP4
GB300 NVL72
$0.52559% cheaper than B200
Dynamo SGLang · FP4
Inference Cost per Million Tokens
Hyperscaler cost ↓ Lower is better Source: InferenceX & SemiAnalysis Market July 2026 AI Cloud TCO Model
| Model · Scenario | B200 · Reference | MI355X | B300 | GB200 NVL72 | GB300 NVL72 |
|---|---|---|---|---|---|
DeepSeek V4 Pro 1.6T8K/1K | vLLM · FP4 | — cannot reach @150 | $0.64250% cheaper than B200 Dynamo SGLang · FP4 | $1.13311% cheaper than B200 Dynamo SGLang · FP4 | $0.52559% cheaper than B200 Dynamo SGLang · FP4 |
DeepSeek V4 Pro 1.6TLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | — cannot reach @150 | — cannot reach @150 | — cannot reach @150 | — no data for this scenario | $0.107No B200 baseline to compare against Dynamo vLLM · FP4 |
Kimi K3 2.8TLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | — cannot reach @150 | — no data for this scenario | $0.229No B200 baseline to compare against vLLM · FP4 | — cannot reach @150 | — no data for this scenario |
MiniMax M3 428B8K/1K | vLLM · FP4 | $0.248145% more expensive than B200 vLLM · FP4 | $0.12321% more expensive than B200 vLLM · FP4 | $0.904791% more expensive than B200 Dynamo vLLM · FP8 · STP | $0.268164% more expensive than B200 Dynamo vLLM · FP8 |
MiniMax M3 428BLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | vLLM · FP4 | — no data for this scenario | $0.01911% cheaper than B200 vLLM · FP4 | — no data for this scenario | — no data for this scenario |
GLM5.2Long Context Multi-Turn Realistic Agentic Scenario (AgentX) | SGLang · FP4 | — no data for this scenario | $0.077About the same cost as B200 SGLang · FP4 | — no data for this scenario | — no data for this scenario |
Qwen3.5 397B8K/1K | SGLang · FP4 | $0.14779% more expensive than B200 SGLang · FP4 | $0.0887% more expensive than B200 SGLang · FP4 | $0.296260% more expensive than B200 Dynamo SGLang · FP8 | $0.335308% more expensive than B200 Dynamo SGLang · FP8 |
Qwen3.5 397BLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | SGLang · FP4 | — no data for this scenario | $0.01559% more expensive than B200 SGLang · FP4 | — no data for this scenario | $0.01116% more expensive than B200 Dynamo SGLang · FP4 |
- View details
DeepSeek V4 Pro 1.6T
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
View detailsB200—cannot reach @150MI355X—cannot reach @150B300—cannot reach @150GB200 NVL72—no data for this scenarioGB300 NVL72$0.107No B200 baseline to compare againstDynamo vLLM · FP4Kimi K3 2.8T
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
View detailsB200—cannot reach @150MI355X—no data for this scenarioB300$0.229No B200 baseline to compare againstvLLM · FP4GB200 NVL72—cannot reach @150GB300 NVL72—no data for this scenarioMiniMax M3 428B
8K/1K
View detailsB200vLLM · FP4MI355X$0.248145% more expensive than B200vLLM · FP4B300$0.12321% more expensive than B200vLLM · FP4GB200 NVL72$0.904791% more expensive than B200Dynamo vLLM · FP8STPGB300 NVL72$0.268164% more expensive than B200Dynamo vLLM · FP8MiniMax M3 428B
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
View detailsB200vLLM · FP4MI355X—no data for this scenarioB300$0.01911% cheaper than B200vLLM · FP4GB200 NVL72—no data for this scenarioGB300 NVL72—no data for this scenarioGLM5.2
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
View detailsB200SGLang · FP4MI355X—no data for this scenarioB300$0.077About the same cost as B200SGLang · FP4GB200 NVL72—no data for this scenarioGB300 NVL72—no data for this scenarioQwen3.5 397B
8K/1K
View detailsB200SGLang · FP4MI355X$0.14779% more expensive than B200SGLang · FP4B300$0.0887% more expensive than B200SGLang · FP4GB200 NVL72$0.296260% more expensive than B200Dynamo SGLang · FP8GB300 NVL72$0.335308% more expensive than B200Dynamo SGLang · FP8Qwen3.5 397B
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
View detailsB200SGLang · FP4MI355X—no data for this scenarioB300$0.01559% more expensive than B200SGLang · FP4GB200 NVL72—no data for this scenarioGB300 NVL72$0.01116% more expensive than B200Dynamo SGLang · FP4
— = no result. ∞ = B200 baseline unavailable.
If a chip does not have FP4 spec decoding available, the next best available configuration is used.