mixtral-8x22b
best known per operation on the selected GPU
4 operations · 1 family · 44 eligible runs
0 of 4 operations have a usable best known on NVIDIA B200 · 4 without one
Best known on NVIDIA B200
Operation · workload
Best known
Record
Trust
Observed
gqa-paged-attention3 operations
GQA paged decode h24 kv4 d128 ps64bf16 · [1, 24, 128] · num pages 425 · num kv indices 361flashinfer / wrapper1b7890not usable359.8µsReported · Apache-2.0 · source2026-03-31stale
No entry in this cohort passes the usability policy (no install recipe); the fastest known is shown. Measured by the source's own harness. Ranked by its metric under ranking-v1.
source mirroredApache-2.0no install recipeCohort on the operation page →View source →Vendor the source →Run detail →
GQA paged prefill causal h24 kv4 d128 ps1bf16 · [154, 24, 128]flashinfer / wrapper0b1670not usable35.8µsReported · Apache-2.0 · source2026-03-31stale
No entry in this cohort passes the usability policy (no install recipe); the fastest known is shown. Measured by the source's own harness. Ranked by its metric under ranking-v1.
source mirroredApache-2.0no install recipeCohort on the operation page →View source →Vendor the source →Run detail →
GQA paged prefill causal h24 kv4 d128 ps64bf16 · [1, 24, 128]flashinfer / wrapperbb278dnot usable21.9µsReported · Apache-2.0 · source2026-03-31stale
No entry in this cohort passes the usability policy (no install recipe); the fastest known is shown. Measured by the source's own harness. Ranked by its metric under ranking-v1. 1 more entry in this cohort.
source mirroredApache-2.0no install recipeCohort on the operation page →View source →Vendor the source →Run detail →
Coverage gaps on NVIDIA B200
Operation
Family
Status
gqa-paged-attention
measured · no usable entry (no install recipe)
gqa-paged-attention
measured · no usable entry (no install recipe)
No eligible evidence yet for these operations. Challenges →
Sources: FlashInfer-Bench (2026-03-31) · Apache-2.0How comparison worksJSON