docs: update benchmark numbers

[github-actions]

Benchmark refresh

Run: 35519999891

Criterion Benchmark Results

Baseline is the per-benchmark median of the last 5 stored runs, so one unusually fast or slow runner cannot skew the comparison. The range column is the spread across those runs.

Benchmark	Baseline (ns/iter)	Range	Current	Change
batch_execute_statement_25	719,195 (n=2)	655,396 - 782,994	415,685	-42.2%
batch_get_item_100	1,404,603	814,524 - 1,451,697	948,986	-32.4%
batch_write_item_25	957,722	533,898 - 1,179,405	675,436	-29.5%
delete_item	46,152	26,334 - 53,068	33,401	-27.6%
get_item	14,424	8,000 - 15,230	9,826	-31.9%
put_item/put_item/large	269,725	115,520 - 278,302	145,467	-46.1%
put_item/put_item/medium	36,564	19,445 - 46,751	25,477	-30.3%
put_item/put_item/small	20,582	10,759 - 26,885	12,922	-37.2%
query_base_table	1,159,394	666,833 - 1,204,548	820,961	-29.2%
query_gsi	22,277	12,359 - 25,843	15,019	-32.6%
scan_with_filter	8,730,220	5,182,664 - 9,113,670	6,103,566	-30.1%
transact_write_items_4	224,380	112,673 - 272,395	157,077	-30.0%
update_item	163,483	85,461 - 170,240	151,778	-7.2%

All benchmarks within 50% of the 5-run median.

Runs in the baseline

• runs/2026-07-24-f7e7d96
• runs/2026-07-30-bc2a16c
• runs/2026-07-30T220642Z-be8bfbc
• runs/2026-08-25T200503Z-54a2cb6
• runs/2026-09-20T153257Z-49c6338

Before merging, eyeball the numbers and re-run this workflow if any look like noise rather than real change.