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 | TRTLLM · FP4 | $0.15943% more expensive than B200 ATOM¹ · FP4 | $0.13420% more expensive than B200 TRTLLM · FP4 | $0.514360% more expensive than B200 Dynamo vLLM · FP4 | $0.21391% more expensive than B200 Dynamo SGLang · FP4 · STP |
DeepSeek V4 Pro 1.6TLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | vLLM · FP4 | $0.01151% more expensive than B200 MoRI SGLang · FP4 | $0.00918% more expensive than B200 vLLM · FP4 | no exact @50 result | no exact @50 result |
Kimi K3 2.8TLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | Dynamo vLLM · FP4 | $0.04736% cheaper than B200 ATOM¹ · FP4 | $0.10340% more expensive than B200 vLLM · FP4 | $0.11353% more expensive than B200 Dynamo vLLM · FP4 | $0.03947% cheaper than B200 Dynamo vLLM · FP4 |
Kimi K2.5/2.6/2.7-Code 1T8K/1K | Dynamo TRTLLM · FP4 · STP | $0.13257% more expensive than B200 ATOM¹ · FP4 · STP | $0.179112% more expensive than B200 vLLM · FP4 · STP | $0.0797% cheaper than B200 Dynamo TRTLLM · FP4 · STP | $0.09817% more expensive than B200 Dynamo TRTLLM · FP4 · STP |
MiniMax M3 428B8K/1K | vLLM · FP8 | $0.05320% cheaper than B200 ATOM¹ · FP4 | $0.05616% cheaper than B200 vLLM · FP4 | $0.10862% more expensive than B200 Dynamo vLLM · FP8 · STP | $0.11065% more expensive than B200 Dynamo vLLM · FP8 · STP |
MiniMax M3 428BLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | no exact @50 result | no exact @50 result | no exact @50 result | no exact @50 result | no exact @50 result |
GLM5.2Long Context Multi-Turn Realistic Agentic Scenario (AgentX) | no exact @50 result | $0.022No B200 baseline to compare against SGLang · FP4 | no exact @50 result | no exact @50 result | no exact @50 result |
Qwen3.5 397B8K/1K | TRTLLM · FP4 | $0.05143% more expensive than B200 SGLang · FP4 | $0.036About the same cost as B200 SGLang · FP4 | no exact @50 result | $0.04011% more expensive than B200 Dynamo TRTLLM · FP4 · STP |
Qwen3.5 397BLong Context Multi-Turn Realistic Agentic Scenario (AgentX) | SGLang · FP8 | $0.00858% cheaper than B200 SGLang · FP8 | $0.01425% cheaper than B200 SGLang · FP8 | $0.00668% cheaper than B200 Dynamo SGLang · FP4 | no exact @50 result |
8K/1K
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
8K/1K
8K/1K
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
8K/1K
Long Context Multi-Turn Realistic Agentic Scenario (AgentX)
— = no result. ∞ = B200 baseline unavailable.
If a chip does not have FP4 spec decoding available, the next best available configuration is used.
