Skip to content
Mock data. Run values, controls, targets and reproductions on this preview are invented for layout. Claim-card text, upstream figures, the reference price and the RFC list come from the repository.
Inference ProofsMenu

Reference system

ref-recompute

Reference row: recomputes the workload's forward pass without proving and compares the output positions (RFC-0015).

Claim card

A reference row: it runs the model without proving and makes no proof claim (RFC-0015).

Rows

One block per workload and configuration. Repetition strips use each metric's own axis.

Workload sash-l31-8b-c4096

ref-recompute on sash-l31-8b-c4096

profile prover-h100tier reported
Measured

whole_wall_seconds

9.827 s

scope whole_wallgrade measuredtier reportedmock

mean 9.827, sd 0.1106, min 9.710, max 9.930, n 3, n_failed 0

mean 9.827 srepetition 1: 9.84 srepetition 2: 9.71 srepetition 3: 9.93 ssource: mock harness record

tokens_covered_total

4,608 token

grade measuredtier reportedmock

mean 4,608, sd 0, min 4,608, max 4,608, n 3, n_failed 0

source: mock harness record
Derived (computed by the derivation code)

rate_whole_wall

469.0 token/s

scope whole_wallgrade derivedtier reportedmock

mean 469.0, sd 5.289, min 464.0, max 474.6, n 3, n_failed 0

mean 469.0 token/srepetition 1: 468.2926829268293 token/srepetition 2: 474.56230690010295 token/srepetition 3: 464.04833836858006 token/ssource: mock derivation record

cost_per_1k_tokens_whole_wall

0.002541 USD/1k_token

scope whole_wallgrade derivedtier reportedmock

mean 0.002541, sd 2.860e-5, min 0.002511, max 0.002568, n 3, n_failed 0

mean 0.002541 USD/1k_tokenrepetition 1: 0.002544704861111111 USD/1k_tokenrepetition 2: 0.002511085792824074 USD/1k_tokenrepetition 3: 0.002567979600694444 USD/1k_tokensource: mock derivation record
Runs
  • mock-rr-c4096-r1completed
  • mock-rr-c4096-r2completed
  • mock-rr-c4096-r3completed
  • Superseded, failed and ineligible runs never contribute to statistics (RFC-0009 S7).

Controls

All controls

Reference rows run no negative controls: they produce no proof.

Reference rows on the same workloads

RFC-0015

ref-serve on sash-l31-8b-c4096

profile prover-h100tier reported
Measured

whole_wall_seconds

3.420 s

scope whole_wallgrade measuredtier reportedmock

mean 3.420, sd 0.03000, min 3.390, max 3.450, n 3, n_failed 0

mean 3.420 srepetition 1: 3.42 srepetition 2: 3.39 srepetition 3: 3.45 ssource: mock harness record

tokens_covered_total

4,608 token

grade measuredtier reportedmock

mean 4,608, sd 0, min 4,608, max 4,608, n 3, n_failed 0

source: mock harness record
Derived (computed by the derivation code)

rate_whole_wall

1,347 token/s

scope whole_wallgrade derivedtier reportedmock

mean 1,347, sd 11.82, min 1,336, max 1,359, n 3, n_failed 0

mean 1,347 token/srepetition 1: 1347.3684210526317 token/srepetition 2: 1359.29203539823 token/srepetition 3: 1335.6521739130435 token/ssource: mock derivation record

cost_per_1k_tokens_whole_wall

0.0008844 USD/1k_token

scope whole_wallgrade derivedtier reportedmock

mean 0.0008844, sd 7.758e-6, min 0.0008767, max 0.0008922, n 3, n_failed 0

mean 0.0008844 USD/1k_tokenrepetition 1: 0.0008844401041666667 USD/1k_tokenrepetition 2: 0.000876681857638889 USD/1k_tokenrepetition 3: 0.0008921983506944446 USD/1k_tokensource: mock derivation record
Runs
  • mock-rs-c4096-r1completed
  • mock-rs-c4096-r2completed
  • mock-rs-c4096-r3completed
  • Superseded, failed and ineligible runs never contribute to statistics (RFC-0009 S7).

Reply from the system maintainers

No reply for this release. System maintainers may write up to 300 words, shown here unedited (RFC-0009 §Right of reply).