gpt-oss 120B — H100 vs MI355X
Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI355X (AMD CDNA 4) on gpt-oss 120B. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
H100 / MI355X on gpt-oss 120B at 91 tok/s/user: 6311 / 26707 tok/s/chip, $0.05 / $0.02 per million tokens. MI355X is 230% cheaper per token; MI355X delivers 323% more tok/s/chip.
Around the middle of the 38–250 tok/s/user interactivity band, at 144 tok/s/user on gpt-oss 120B: H100 runs 4116 tok/s/chip at $0.08/M tokens, MI355X runs 16480 at $0.03/M. MI355X is 212% cheaper per token; MI355X delivers 300% more tok/s/chip.
Setting 197 tok/s/user as the target on gpt-oss 120B, H100 produces 2636 tok/s/chip ($0.12 per million tokens) and MI355X produces 10740 ($0.04). MI355X is 218% cheaper per token; MI355X delivers 307% more tok/s/chip. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:6310.8MI355X:26707.1 | H100:4115.7MI355X:16480.1 | H100:2636.0MI355X:10739.7 |
| Cost ($/M tok) | H100:$0.051MI355X:$0.016 | H100:$0.079MI355X:$0.025 | H100:$0.123MI355X:$0.039 |
| tok/s/MW | H100:4606456MI355X:12778496 | H100:3004142MI355X:7885227 | H100:1924056MI355X:5138603 |
| Concurrency | H100:~32MI355X:~34 | H100:~13MI355X:~13 | H100:~6MI355X:~6 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
Total Tokens per $1 USD (Owning - Hyperscaler) vs. Interactivity
gpt-oss 120B • FP4 • 8K / 1K • Source: SemiAnalysis InferenceX™
TCO $/chip/hr: H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68
Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate cost per million tokens per decode chip or per prefill chip, rather than per total chip count. This makes direct token cost comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate input throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct input throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate output throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct output throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate power per decode chip or per prefill chip, rather than per total chip count. This makes direct power comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate Joules per decode chip or per prefill chip, rather than per total chip count. This makes direct Joules per token comparison with aggregated configs not an apples-to-apples comparison.
No data available
Please change the model, sequence, precision, date range or chip selection.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip