MiniMax M3 428B · Chip comparison
MiniMax M3 428B — B200 vs GB200 NVL72
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) on MiniMax M3 428B. 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.
AgentX replays real coding-agent sessions rather than fixed-length prompts, so context grows turn over turn and most of each request is served from cache instead of being recomputed. That turns the comparison into a systems question: KV transfer between nodes, prefix-aware routing, and cache capacity all move the curve alongside raw chip throughput. Fixed-sequence workloads stay the clean baseline for kernel and silicon performance, so the two scenarios answer different questions about the same hardware. Learn more about AgentX →
Near the low end of the 83–423 tok/s/user interactivity band, at 168 tok/s/user on MiniMax M3 428B: B200 runs 33,580 tok/s/chip at $0.014 per million tokens, GB200 NVL72 runs 33,531 at $0.015. B200 is 7% cheaper per token; throughput per chip is essentially tied.
Setting 253 tok/s/user as the target on MiniMax M3 428B, B200 produces 31,374 tok/s/chip ($0.015 per million tokens) and GB200 NVL72 produces 16,343 ($0.032). B200 is 52% cheaper per token; B200 delivers 92% more tok/s/chip.
At 338 tok/s/user interactivity on MiniMax M3 428B, B200 delivers 23,876 tok/s/chip at $0.020 per million tokens; GB200 NVL72 delivers 8,176 tok/s/chip at $0.063. B200 is 68% cheaper per token; B200 delivers 192% more tok/s/chip at this point. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
GB200 NVL72 vs B200: full specs & benchmarks across models →
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B20033,580GB200 NVL7233,531 | B20031,374GB200 NVL7216,343 | B20023,876GB200 NVL728,176 |
| Cost ($/M tok) | B200$0.014GB200 NVL72$0.015 | B200$0.015GB200 NVL72$0.032 | B200$0.020GB200 NVL72$0.063 |
| tok/s/MW | B20019,637,248GB200 NVL7217,930,819 | B20018,347,533GB200 NVL728,739,563 | B20013,962,707GB200 NVL724,372,272 |
| Concurrency | B200~20GB200 NVL72~19 | B200~20GB200 NVL72~9 | B200~16GB200 NVL72~2 |