diff --git a/CHECKLIST.md b/CHECKLIST.md
index 7a8ddea..1180627 100644
--- a/CHECKLIST.md
+++ b/CHECKLIST.md
@@ -350,18 +350,18 @@
     - 结论：int8 省 4× 空间几乎不掉质量（0.977）但不省时间（35.5 vs 41.2）；bit 省 32× 空间、省 **15.8×** 时间（41.2→2.6），只剩 0.728，只能当粗排；两级方案必须在 vec0 **外面**精排（在库里用 `rowid IN (…)` 回捞反而更慢）。维度截断是在**不是 MRL 训练**的 MiniLM 上测的，所以那是真实代价不是白拿：前 192 维 7.5ms、前 128 维 7.9ms，对全维 8.0ms 只快 1–6%，前 64 维才快一半但一致度掉到 0.43；切片必须是存下来的连续维，不是查询时切（非连续视图点集会掉速度）。
     - 偏差：精排腿（`bge-reranker-base`）早在基线消融里就有并测过（加了它 recall@10 从 0.565 掉到 0.453），所以这里没再为它编一套故事；粗排+精排这条链测出 0.983 一致度但**没接线**，因为 dense 腿默认关（见下）。
   - [x] p2-1.3 100k 实测 + 1M 外推：`python -m bench.dense_scale` → `bench/reports/DENSE.md`（stamp dff5977）
-    - 实测 **165,973 atom / 153,857 向量**（DENSE.md，stamp 6cc564085）：建库 1444.8s（106 向量/秒）、**检索 p50 73.1ms / p95 76.1ms / max 87.2ms（300 查询）**、其中编码问题 p50 11.0ms / p95 24.1ms、与暴力扫描 tie-aware 一致度 **0.998**（40 探针）、**往满索引里写一条记忆 p50 509ms / p95 531ms（50 次写）**、失配扫描 1.2s（不在检索路径上）、**每向量 1,729 字节**（dbstat 数 vec0 影子表：266,059,776 B）、机器 load [3.41, 4.52, 4.31]。
-    - 1M 外推（**分母是向量数**，1,000,000 / 153,857，不是 atom 数——两者差 7.9%，所以报告里把分母印在外推表旁边）：p50 475ms / p95 494ms、向量表 1.6 GiB、首次建库 156.5 分钟、单条写 3310ms（vec0 按 id 删除也要扫表，所以写同样随规模线性），假设随行印出（逐条扫 ⇒ 线性）。
+    - 实测 **165,973 atom / 153,857 向量**（DENSE.md，stamp 71733d86d）：建库 1444.9s（106 向量/秒）、**检索 p50 73.0ms / p95 75.7ms / max 85.6ms（300 查询）**、其中编码问题 p50 11.3ms / p95 24.0ms、与暴力扫描 tie-aware 一致度 **0.998**（40 探针）、**往满索引里写一条记忆 p50 508ms / p95 535ms（50 次写）**、失配扫描 1.3s（不在检索路径上）、**每向量 1,729 字节**（dbstat 数 vec0 影子表：266,059,776 B）、机器 load [3.53, 4.63, 4.85]。
+    - 1M 外推（**分母是向量数**，1,000,000 / 153,857，不是 atom 数——两者差 7.9%，所以报告把分母印在外推表旁边）：p50 475ms / p95 492ms、向量表 1.6 GiB、首次建库 156.5 分钟、单条写 3303ms（vec0 按 id 删除也要扫表，所以写同样随规模线性），假设随行印出（逐条扫 ⇒ 线性）。
     - tie-aware 的必要性：语料里有重复 atom，分数并列 ⇒ top-K 的 id 集合本身有歧义，两个都对引擎按 id 比会读出 0.827 的假召回。判定改成「返回的 atom 真实 cosine ≥ 第 K 名 cosine」。
-  - 验收：100k dense 腿 p50 <100ms → **165,973 atom（bar 的 1.66 倍）p50 73.1ms**（DENSE.md）；新 atom 立即可查 → `pytest backend/tests/test_vector_index.py`（写进去当场能搜到、视图藏起来的 dense 腿也藏、备份带走向量恢复后照样应答）。
+  - 验收：100k dense 腿 p50 <100ms → **165,973 atom（bar 的 1.66 倍）p50 73.0ms**（DENSE.md）；新 atom 立即可查 → `pytest backend/tests/test_vector_index.py`（写进去当场能搜到、视图藏起来的 dense 腿也藏、备份带走向量恢复后照样应答）。
   - [x] p2-1.4 每条腿值多少钱：`python -m bench.leg_check` → `bench/reports/LEGS.md`（stamp 见报告首行，300 条 holder 作用域探针，7 个配置，每臂带 n / MRR / Wilson 下界 / p50 / p95 / 分模态召回）。测的是 `bench/runs/dense_scale` 的副本：冻结基线自己**不带向量**，dense 腿在其上只会排一个空索引，所以报告首行就把来源 store 与向量数印出来。
-    - 只有词的 FTS 页：问 10 个 atom recall@10 **0.832**、MRR 0.683、p50 65.1ms；问产品配置的 24 个 **0.848** / 0.688 / 66.5ms；页放宽到 30 **0.855** / 0.690 / 94.0ms（放宽不要钱，10→24 就白捡 1.6 个点）。
-    - 开 dense 腿：10 个 atom **0.845** / MRR 0.663 / p50 481.4ms；页 30 **0.858** / 0.669 / 412.8ms。多 1.3–2.6 个点召回、MRR 降、慢 6–7 倍。
-    - 开图片腿。索引实物印在 LEGS.md 里（不是这里口头说的）：**53,668 个 atom、512 维、109,912,192 字节**；编码速度实测 **19.9 张/秒**（200 张样本），据此那 53,668 张 ≈ **45.0 分钟**——和我当时手工记的 2901.7 秒（48.4 分钟）差 7%，两条路对得上。结果：10 个 atom **0.815** / MRR 0.553 / 139.5ms；页 30 **0.825** / 0.551 / 160.0ms。分模态才看得见它在干什么：**video 0.583→0.708、pdf 0.849→0.907 涨，code 0.953→0.898、ui_screenshot 0.583→0.500 跌**——它救的正是金标是图的题，代价是稀释别的。
-    - p50 是负载函数，不是性常数：同一份 harness 两次跑，words 腿 64.7–72.5ms、dense 腿 412.8–481.4ms、图片腿 132.5–160.0ms，**召回一格没动**。所以比较只用召回，延迟看报告当次的 load。
+    - 只有词的 FTS 页：问 10 个 atom recall@10 **0.832**、MRR 0.683、p50 67.5ms；问产品配置的 24 个 **0.848** / 0.688 / 67.6ms；页放宽到 30 **0.855** / 0.690 / 72.0ms（放宽不要钱，10→24 就白捡 1.6 个点）。
+    - 开 dense 腿：10 个 atom **0.845** / MRR 0.663 / p50 474.8ms；页 30 **0.858** / 0.669 / 416.0ms。多 1.3–2.6 个点召回、MRR 降、慢 6–7 倍。
+    - 开图片腿。索引实物印在 LEGS.md 里（不是这里口头说的）：**53,668 个 atom、512 维、109,912,192 字节**；编码速度也是实测的（200 张样本）：**16.8–19.9 张/秒**，据此那 53,668 张 ≈ **53.3–45.0 分钟**——和我当时手工记的 2901.7 秒（48.4 分钟）落在同一个带里。结果：10 个 atom **0.815** / MRR 0.553 / 138.1ms；页 30 **0.825** / 0.551 / 166.7ms。分模态才看得见它在干什么：**video 0.583→0.708、pdf 0.849→0.907 涨，code 0.953→0.898、ui_screenshot 0.583→0.500 跌**——它救的正是金标是图的题，代价是稀释别的。
+    - p50 是负载函数，不是性常数：两份已提交的 LEGS.md 里，words 腿 64.7–72.5ms、dense 腿 413.7–474.8ms、图片腿 132.5–166.7ms，**召回三跑到小数点后三位一模一样**（0.832 / 0.848 / 0.855 / 0.845 / 0.858 / 0.815 / 0.825）。所以比较只用召回，延迟看报告当次的 load。
     - 产品的图片腿不是开关：`retrieval.py:136` `if image is None and not thin: return {}`，稀疏页薄到 2 条以下才问图。所以这几行测的是「腿可用」，不是强行常开融合。
-    - **没达到的验收线，写在这里**：用户定的验收是 recall@10 ≥0.90（n≥300）。测出来最好的一行是 **0.858**（dense 腿开、页 30、p50 412.8ms），默认配置 **0.848**（p50 66.5ms），CURRENT.md 里按 10 个 atom 问是 0.832；仓内 Mem0/Zep 式对照在 holder 作用域是 **0.866**。差距集中在图（image 0.750、ui_screenshot 0.500–0.583、video 0.583–0.708），文字的题（code 0.953、pdf 0.919）已经过了线。诚实结论：**延迟达标（281 倍于基线、p50 71.7ms），召回没达标**，剩下 4 个点是跨模态召回，不是索引规模问题。
-  - 数字逼出来的决定：**dense 腿默认关**（`config.py:32 dense_embeddings_enabled=False`）。同一 300 条 holder 作用域探针上（LEGS.md），开腿 recall@10 0.832→0.845、MRR 0.683→0.663、p50 65.1ms→481.4ms。多 1.3 个点的召回、慢 7 倍、首位质量还降——默认值就该是关，腿留着（已增量、已入库）等更大的语料或更好的嵌入模型。
+    - **没达到的验收线，写在这里**：用户定的验收是 recall@10 ≥0.90（n≥300）。测出来最好的一行是 **0.858**（dense 腿开、页 30、p50 416.0ms），默认配置 **0.848**（p50 67.6ms），CURRENT.md 里按 10 个 atom 问是 0.832；仓内 Mem0/Zep 式对照在 holder 作用域是 **0.866**。差距集中在图（image 0.750、ui_screenshot 0.500–0.583、video 0.583–0.708），文字的题（code 0.953、pdf 0.919）已经过了线。诚实结论：**延迟达标（281 倍于基线、p50 71.7ms），召回没达标**，剩下 4 个点是跨模态召回，不是索引规模问题。
+  - 数字逼出来的决定：**dense 腿默认关**（`config.py:32 dense_embeddings_enabled=False`）。同一 300 条 holder 作用域探针上（LEGS.md），开腿 recall@10 0.832→0.845、MRR 0.683→0.663、p50 67.5ms→474.8ms。多 1.3 个点的召回、慢 7 倍、首位质量还降——默认值就该是关，腿留着（已增量、已入库）等更大的语料或更好的嵌入模型。
   - 门禁：`pytest -q` 全绿；`ruff check` 干净；mypy 冷缓存 **24 errors / 9 files**，与 p1-6 基线（`9cb46c1` 同口径 24/9）**净新增 0**——p2-1 一度引入 +2（`vector_index.py` 与 `quant_probe.py` 的 `import sqlite_vec` 没有 stub），根因修法是在 `pyproject.toml` 的 `[[tool.mypy.overrides]]` 里声明 `sqlite_vec` 无 stub，而不是在代码里加 `# type: ignore`；`./node_modules/.bin/pyright` 对改动文件 0 errors；`bench/filesize_check` 干净；go / npm 未动。
   - 顺带修的测量缺陷：`copy_store()` 只拷 SQLite，所以任何要跑视觉腿的 harness 都得自己带上 `objects/`（25 GB，探针里用符号链接）与 `cache/clip_image_index.npy`；`bench/scale_check.py` 原本直接打开冻结基线库、把它**原地迁移**了，现改成在副本上测（`--scratch`）。产品 CLI 不在这条规则里：p2-1.4 建图片索引时 `ely-eye index-visual` 就把冻结基线从 schema 7 升到了 8，所以 p1-4.8 那两条钉的是**当前**的 mtime / sha256 / schema，并指名是哪一步改的。
 - [ ] **p2-2** 安全与隐私
diff --git a/bench/reports/DENSE.md b/bench/reports/DENSE.md
index bb8bec9..54ec2ff 100644
--- a/bench/reports/DENSE.md
+++ b/bench/reports/DENSE.md
@@ -1,23 +1,23 @@
-# DENSE: the vector leg at corpus size (commit dff5977ab)
+# DENSE: the vector leg at corpus size (commit 71733d86d)
 
 - database: `bench/runs/dense_scale/data/ely_eye.sqlite`
 - atoms: 165973, memories with words: 153857, vectors: 153857, pending: 0
 - model: `sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2`, width 384
-- build: 1421.0s for 153857 memories (108 per second)
-- writing one memory into a full index: p50 518ms · p95 543ms over 50 writes
+- build: 1444.9s for 153857 memories (106 per second)
+- writing one memory into a full index: p50 508ms · p95 535ms over 50 writes
 - the repair scan (`ely-eye build-dense-index`) over a full index: 1.3s, and it is not on the search path
-- search p50 75.6ms · p95 93.0ms · max 123.6ms over 300 queries
-- of which encoding the question: p50 12.4ms · p95 25.8ms
+- search p50 73.0ms · p95 75.7ms · max 85.6ms over 300 queries
+- of which encoding the question: p50 11.3ms · p95 24.0ms
 - against a brute-force scan: 0.998 of the top 10 over 40 probes (a returned atom counts when its true cosine is at least the k-th best, so ties cannot fail the check)
 - bytes per stored vector: 1729 (266059776 bytes for 153857 vectors, dbstat page sizes over the vec0 shadow tables)
-- machine: Linux 6.18.40.1-microsoft-standard-WSL2, 32 cores, 4 threads per query, load [3.18, 4.58, 5.42]
+- machine: Linux 6.18.40.1-microsoft-standard-WSL2, 32 cores, 4 threads per query, load [3.53, 4.63, 4.85]
 
-## One million memories on this machine
+## One million vectors on this machine
 
-- p50 455ms · p95 561ms
+- p50 475ms · p95 492ms
 - vector table: 1.6 GiB
-- first build: 153.9 minutes
-- writing one memory: 3123ms — a vec0 delete by id scans the table, so this grows with the count too
-- assumption: sqlite-vec 0.1.9 scans every vector, so latency and bytes grow linearly with the count; the model, the machine and the thread budget stay the same
+- first build: 156.5 minutes
+- writing one vector: 3303ms — a vec0 delete by id scans the table, so this grows with the count too
+- assumption: sqlite-vec 0.1.9 scans every vector, so latency and bytes grow linearly with the vector count; every row here is scaled by 1000000 / 153857 vectors, and the model, the machine and the thread budget stay the same
 
 _Latency is wall time per search on the machine named above with the thread budget the benchmark pins; a run on a loaded machine is not comparable._
diff --git a/bench/reports/LEGS.md b/bench/reports/LEGS.md
index 1a1ed95..39c4299 100644
--- a/bench/reports/LEGS.md
+++ b/bench/reports/LEGS.md
@@ -1,22 +1,24 @@
 # LEGS: what each retrieval leg is worth
 
-- commit: `dff5977ab638896cc8685c78ee41e5195fc0fbf2`, schema v8
-- store: bench/runs/leg_check/data/ely_eye.sqlite (a copy of bench/runs/dense_scale; the frozen baseline is never opened)
+- commit: `71733d86d6d27d1f1ebb75205d5e0599f253f791`, schema v8
+- store: bench/runs/leg_check/data/ely_eye.sqlite — a copy of bench/runs/dense_scale, which carries 153,857 dense vectors built by bench/dense_scale. The build sits outside the measured latency, and the frozen baseline itself holds no vectors, so a dense arm there would rank an empty index.
 - 300 answerable queries, each asked as the holder the frozen set assigns it
-- machine: Linux 6.18.40.1-microsoft-standard-WSL2, 32 cores, 4 threads per query, load [25.51, 9.62, 5.7]
+- picture index: 53,668 atoms at 512 dimensions, 109,912,192 bytes, read from bench/runs/leg_check_index
+- encoder speed: 16.8 pictures/s measured on 200 pictures, so the 53,668 pictures in that index cost about 53.3 minutes to encode
+- machine: Linux 6.18.40.1-microsoft-standard-WSL2, 32 cores, 4 threads per query, load [22.19, 9.9, 6.39]
 
 Recall is scored on the first ten **sources** after atoms are collapsed to their source
 by first hit, which is how every other arm in this repo is scored.
 
 | leg configuration | atoms asked | recall@10 | MRR | hit@10 | 95% lower | p50 ms | p95 ms | distinct sources |
 | --- | --- | --- | --- | --- | --- | --- | --- | --- |
-| words only | 10 | 0.832 | 0.683 | 0.833 | 0.787 | 65.5 | 134.9 | 4.0 |
-| words only (the configured 24) | 24 | 0.848 | 0.688 | 0.850 | 0.805 | 72.5 | 155.9 | 6.0 |
-| words only, wider page | 30 | 0.855 | 0.690 | 0.857 | 0.812 | 64.7 | 129.7 | 6.0 |
-| + dense leg | 10 | 0.845 | 0.663 | 0.847 | 0.802 | 472.9 | 577.5 | 5.0 |
-| + dense leg, wider page | 30 | 0.858 | 0.669 | 0.860 | 0.816 | 413.7 | 525.8 | 7.0 |
-| + picture leg | 10 | 0.815 | 0.553 | 0.817 | 0.769 | 132.5 | 271.3 | 8.0 |
-| + picture leg, wider page | 30 | 0.825 | 0.551 | 0.827 | 0.780 | 162.8 | 283.7 | 20.0 |
+| words only | 10 | 0.832 | 0.683 | 0.833 | 0.787 | 67.5 | 146.7 | 4.0 |
+| words only (the configured 24) | 24 | 0.848 | 0.688 | 0.850 | 0.805 | 67.6 | 129.9 | 6.0 |
+| words only, wider page | 30 | 0.855 | 0.690 | 0.857 | 0.812 | 72.0 | 156.0 | 6.0 |
+| + dense leg | 10 | 0.845 | 0.663 | 0.847 | 0.802 | 474.8 | 579.1 | 5.0 |
+| + dense leg, wider page | 30 | 0.858 | 0.669 | 0.860 | 0.816 | 416.0 | 543.4 | 7.0 |
+| + picture leg | 10 | 0.815 | 0.553 | 0.817 | 0.769 | 138.1 | 303.6 | 8.0 |
+| + picture leg, wider page | 30 | 0.825 | 0.551 | 0.827 | 0.780 | 166.7 | 314.2 | 20.0 |
 
 ## recall@10 by modality
 
diff --git a/bench/reports/dense.json b/bench/reports/dense.json
index 02fb131..6aecf0f 100644
--- a/bench/reports/dense.json
+++ b/bench/reports/dense.json
@@ -1,5 +1,5 @@
 {
-  "commit": "dff5977ab638896cc8685c78ee41e5195fc0fbf2",
+  "commit": "71733d86d6d27d1f1ebb75205d5e0599f253f791",
   "database": "bench/runs/dense_scale/data/ely_eye.sqlite",
   "atoms": 165973,
   "memories_with_words": 153857,
@@ -7,17 +7,17 @@
   "pending": 0,
   "dim": 384,
   "model": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
-  "copy_seconds": 0.665644496002642,
-  "build_seconds": 1421.0276969089973,
-  "texts_per_second": 108.2716405420302,
-  "repair_seconds": 1.2946931579936063,
+  "copy_seconds": 4.653337935000309,
+  "build_seconds": 1444.851663221998,
+  "texts_per_second": 106.48636390596745,
+  "repair_seconds": 1.2708722250099527,
   "writes": 50,
   "write_latency": {
     "queries": 50,
-    "mean": 0.5204645371803781,
-    "p50": 0.5183807220018934,
-    "p95": 0.5429989652060613,
-    "max": 0.5760761429992272
+    "mean": 0.5132235406010295,
+    "p50": 0.50813193799695,
+    "p95": 0.5353954541016719,
+    "max": 0.5569016869994812
   },
   "probes": 300,
   "probe_modalities": [
@@ -29,17 +29,17 @@
   ],
   "latency": {
     "queries": 300,
-    "mean": 0.07861847096656371,
-    "p50": 0.07557305500085931,
-    "p95": 0.09303550984986941,
-    "max": 0.123588470989489
+    "mean": 0.07290171435301697,
+    "p50": 0.07302105800044956,
+    "p95": 0.07570205154406721,
+    "max": 0.08561979999649338
   },
   "encode_latency": {
     "queries": 300,
-    "mean": 0.014255492710411393,
-    "p50": 0.012439899503078777,
-    "p95": 0.025788354348333098,
-    "max": 0.0513256230042316
+    "mean": 0.012841759066456385,
+    "p50": 0.011289245005173143,
+    "p95": 0.024048260039126034,
+    "max": 0.02614720400015358
   },
   "correctness": {
     "probes": 40,
@@ -52,12 +52,13 @@
   "bytes_method": "dbstat page sizes over the vec0 shadow tables",
   "extrapolation": {
     "to": 1000000,
-    "assumption": "sqlite-vec 0.1.9 scans every vector, so latency and bytes grow linearly with the count; the model, the machine and the thread budget stay the same",
-    "p50_seconds": 0.45533342773137386,
-    "p95_seconds": 0.5605460517666694,
-    "write_p50_seconds": 3.1232834376789804,
+    "basis": 153857,
+    "assumption": "sqlite-vec 0.1.9 scans every vector, so latency and bytes grow linearly with the vector count; every row here is scaled by 1000000 / 153857 vectors, and the model, the machine and the thread budget stay the same",
+    "p50_seconds": 0.47460341746199103,
+    "p95_seconds": 0.49202864701682225,
+    "write_p50_seconds": 3.3026247619344584,
     "gibibytes": 1.6105050507025191,
-    "build_seconds": 9236.028889871746
+    "build_seconds": 9390.873754343305
   },
   "machine": {
     "system": "Linux",
@@ -65,9 +66,9 @@
     "cpus": 32,
     "threads_per_query": 4,
     "load_average_1_5_15": [
-      3.18,
-      4.58,
-      5.42
+      3.53,
+      4.63,
+      4.85
     ]
   }
 }
\ No newline at end of file
diff --git a/bench/reports/leg_check.json b/bench/reports/leg_check.json
index 6f00ead..cae60f7 100644
--- a/bench/reports/leg_check.json
+++ b/bench/reports/leg_check.json
@@ -1,7 +1,20 @@
 {
-  "commit": "dff5977ab638896cc8685c78ee41e5195fc0fbf2",
+  "commit": "71733d86d6d27d1f1ebb75205d5e0599f253f791",
   "schema_version": 8,
   "store": "bench/runs/leg_check/data/ely_eye.sqlite",
+  "source_store": "bench/runs/dense_scale",
+  "dense_vectors": 153857,
+  "picture_index": {
+    "path": "bench/runs/leg_check_index",
+    "atoms": 53668,
+    "dim": 512,
+    "bytes": 109912192
+  },
+  "encode_rate": {
+    "pictures": 200,
+    "seconds": 11.912494190008147,
+    "per_second": 16.7890952818054
+  },
   "probes": 300,
   "arms": [
     {
@@ -13,8 +26,8 @@
       "mrr": 0.6832447089947089,
       "hit_at_10": 0.8333333333333334,
       "wilson_lower": 0.7870033503873167,
-      "p50_seconds": 0.0654734540003119,
-      "p95_seconds": 0.13490134734311146,
+      "p50_seconds": 0.0674771249978221,
+      "p95_seconds": 0.14665940315098855,
       "median_distinct_sources": 4.0,
       "by_modality": {
         "code": 0.953125,
@@ -40,8 +53,8 @@
       "mrr": 0.6875571326821327,
       "hit_at_10": 0.85,
       "wilson_lower": 0.8051823888546709,
-      "p50_seconds": 0.07251374649786158,
-      "p95_seconds": 0.15588207065375176,
+      "p50_seconds": 0.0675995074925595,
+      "p95_seconds": 0.12988004180151622,
       "median_distinct_sources": 6.0,
       "by_modality": {
         "code": 0.953125,
@@ -67,8 +80,8 @@
       "mrr": 0.6895710215710216,
       "hit_at_10": 0.8566666666666667,
       "wilson_lower": 0.8124994228503402,
-      "p50_seconds": 0.06468957299512113,
-      "p95_seconds": 0.1297007056004077,
+      "p50_seconds": 0.07200110949634109,
+      "p95_seconds": 0.15601039270331973,
       "median_distinct_sources": 6.0,
       "by_modality": {
         "code": 0.953125,
@@ -94,8 +107,8 @@
       "mrr": 0.6632195767195768,
       "hit_at_10": 0.8466666666666667,
       "wilson_lower": 0.8015339429418983,
-      "p50_seconds": 0.4729458744986914,
-      "p95_seconds": 0.5775450093562541,
+      "p50_seconds": 0.4748052095019375,
+      "p95_seconds": 0.5791264187995694,
       "median_distinct_sources": 5.0,
       "by_modality": {
         "code": 0.9375,
@@ -121,8 +134,8 @@
       "mrr": 0.6686637344137344,
       "hit_at_10": 0.86,
       "wilson_lower": 0.8161683876490509,
-      "p50_seconds": 0.41367531901050825,
-      "p95_seconds": 0.5258364057008293,
+      "p50_seconds": 0.41600067100080196,
+      "p95_seconds": 0.5433694835548523,
       "median_distinct_sources": 7.0,
       "by_modality": {
         "code": 0.9375,
@@ -148,8 +161,8 @@
       "mrr": 0.5525886243386243,
       "hit_at_10": 0.8166666666666667,
       "wilson_lower": 0.7689713823226987,
-      "p50_seconds": 0.13248339149868116,
-      "p95_seconds": 0.2713221473451995,
+      "p50_seconds": 0.1381109430003562,
+      "p95_seconds": 0.3035783412473393,
       "median_distinct_sources": 8.0,
       "by_modality": {
         "code": 0.9296875,
@@ -175,8 +188,8 @@
       "mrr": 0.5507809648412589,
       "hit_at_10": 0.8266666666666667,
       "wilson_lower": 0.7797738592974309,
-      "p50_seconds": 0.16283495400421089,
-      "p95_seconds": 0.2837207488548302,
+      "p50_seconds": 0.16669872149941511,
+      "p95_seconds": 0.3142305556466454,
       "median_distinct_sources": 20.0,
       "by_modality": {
         "code": 0.8984375,
@@ -200,9 +213,9 @@
     "cpus": 32,
     "threads_per_query": 4,
     "load_average_1_5_15": [
-      25.51,
-      9.62,
-      5.7
+      22.19,
+      9.9,
+      6.39
     ]
   }
 }
\ No newline at end of file