Cohere-1M — 2026-08-29

c8g.large · peak recall 0.9988 @ 68 QPS

QPS at recall

Recall	Meerkat	pgvector HNSW (halfvec)	Δ vs best competitor
0.90	504	13	+3819%
0.95	338	8	+4044%
0.99	115	3	+4381%

Latency at recall

	Meerkat	pgvector HNSW (halfvec)	Δ vs best competitor
@ recall 0.90
p50	1.75 ms	75.73 ms	-98%
p95	4.31 ms	117.71 ms	-96%
p99	6.18 ms	143.55 ms	-96%
@ recall 0.95
p50	2.65 ms	121.81 ms	-98%
p95	5.35 ms	177.45 ms	-97%
p99	6.73 ms	209.30 ms	-97%
@ recall 0.99
p50	8.62 ms	391.74 ms	-98%
p95	11.39 ms	534.32 ms	-98%
p99	13.09 ms	583.33 ms	-98%

All values are taken at exactly the stated recall: each configuration's own sweep is log-interpolated between its two surrounding measured points (measured values are used verbatim on an exact hit) and the best configuration at that recall is shown — values are never interpolated across different configurations. “—” means no single sweep spans that recall; nothing is extrapolated. Recall is distance-based (ann-benchmarks, tie-aware). Insert/update/delete workloads are not measured.

Interpolated on config — Meerkat @0.90: mkt-rerank20 · pgvector HNSW (halfvec) @0.90: hnsw-m24-halfvec · Meerkat @0.95: mkt-rerank20 · pgvector HNSW (halfvec) @0.95: hnsw-m24-halfvec · Meerkat @0.99: mkt-rerank40 · pgvector HNSW (halfvec) @0.99: hnsw-m24-halfvec

Build comparison

	Meerkat	pgvector HNSW (halfvec)	Δ vs best competitor
Build time	—	107472 s	—
Index size	—	1.9 GB	—

Build stats are for the configuration backing each column's best operating point: pgvector HNSW (halfvec) = hnsw-m24-halfvec, prewarmed.

Plots

Pareto svg · Distance-recall svg · Per-config svg · Build times svg · Index sizes svg

Host

Instance type	c8g.large
Hostname	ip-172-31-37-90
CPU	unknown (2 vCPU)
Memory	4 GB
OS	Ubuntu 26.04 LTS
Kernel	7.0.0-1011-aws

Versions

rekall	0.1.0 (1834b41 / main)
PostgreSQL	PostgreSQL 18.6 (Ubuntu 18.6-1.pgdg26.04+2) on aarch64-unknown-linux-gnu, compiled by gcc (Ubuntu 15.2.0-16ubuntu1) 15.2.0, 64-bit
meerkat	0.1.0 (f9b29b8)
pg_prewarm	1.2
vector	0.8.6

Search

k	10
runs	3
warmup	500
num_queries	10000
S3 archive	s3://rekall-bench/runs/cohere-1m/2026-08-29/c8g.large-073532/

Index builds

Index	Type	Build time	Size	Workers
hnsw-halfvec	pgvector-hnsw	78044.8s	1.9 GB	1
hnsw-m24-halfvec	pgvector-hnsw	107471.7s	1.9 GB	1
mkt-defaults	meerkat	185.5s	266.8 MB	1

Index parameters

hnsw-halfvec: {"ef_construction":200,"encoding":"halfvec","m":16}
hnsw-m24-halfvec: {"ef_construction":200,"encoding":"halfvec","m":24}
mkt-defaults: {"boundary_epsilon":null,"centroid_compression":null,"centroid_fastscan":null,"fan_out":null,"fastscan":null,"fastscan_bits":null,"nlist":null,"rerank_vectors":null,"soar_lambda":null}

Instance settings (observed)

instance type          c8g.large
vCPU (observed)        2
memory (observed)      3.7 GiB
cpu model              unknown
kernel                 7.0.0-1011-aws
data directory         /var/lib/postgresql/18/main
mount point            /
filesystem             ext4
media                  unknown
device                 root
provisioned IOPS       — not an EBS volume

PostgreSQL settings

checkpoint_completion_target = 0.9
cluster_name = 18/main
data_directory = /var/lib/postgresql/18/main
effective_cache_size = 3 GB
effective_io_concurrency = 200
huge_pages = try
jit = off
maintenance_io_concurrency = 200
maintenance_work_mem = 1 GB
max_connections = 100
max_parallel_maintenance_workers = 1
max_parallel_workers = 2
max_wal_size = 16 GB
random_page_cost = 1.1
server_version = 18.6 (Ubuntu 18.6-1.pgdg26.04+2)
shared_buffers = 1 GB
wal_level = replica
work_mem = 32 MB