Model monitor
Google Gemini 3.5 Flash
google/gemini-3.5-flash · via openrouter
As of 2026-08-05: no alarm in 5 of 5 task families · 67 complete panels behind the record.
Served route, trailing 30 days: median latency 4.3 s · p95 9.3 s across 705 probe calls · served by Google (705), unrecorded (15). Observed at panel time — a few calls a day, not continuous uptime.
Certified claim, per family: H0: E[panel_mean_t | prior completed panels] ≥ frozen_null_floor. Fixed panel composition does not assume or establish temporal independence. Recent panels below link to their hash-committed evidence.
Task family
Structured output
- Frozen null floor
- 0.700
- Anytime level α
- 0.050
- Completed panels
- 3
- CS lower
- 0.000
- CS upper
- 1.000
Panel record — 3 completed panels
| Cycle | Completed | Member scores | Mean | Evidence |
|---|---|---|---|---|
| 45 | 2026-08-05 15:05 UTC | 1.001.001.000.381.00 | 0.875 | 8be79ab7618a… |
| 44 | 2026-08-05 14:01 UTC | 1.001.001.001.001.00 | 1.000 | 92288a36f61c… |
| 43 | 2026-08-05 12:59 UTC | 1.001.001.001.001.00 | 1.000 | 55e1264f35b0… |
Newest first. A failed member scores 0 and stays on the record. Each evidence link opens the panel’s full hash-committed record.
Epoch & design record
- Current epoch
03bf900d-503e-4cae-b5c1-c868f411f393- Opened
- 2026-08-05 12:27 UTC
- Previous epoch
9f8af8e9-4c0d-419a-a3a5-146898049eae- Transition
- checker_refresh
- Design hash
8a4f5431d708d308ccd859fc31d097a5db0297535ace2795cab39b5f087eb318- Panel hash
a37ad21aa4b8c3e16dbae958c68a70064a9d95487e131d832f2c601f7fdacf14- Pool hash
8e9496011afbb949e38b3facc664e6dd560f1c5d2155fd424fcb720d084c69ee- Panel size
- 5
- Terminal failure
- zero — a failed member scores 0 and stays on the record
- Pool version
structured-output-v3- Checker version
structured_output-v2- Reference mean μ₀
- 0.700
- Minimum effect δ
- 0.000
- Evidence log e
- -0.23 · max 0.00 · alarm at 3.00 (= ln 1/α)
- Raw scores kept
- 15
- Provider spend, this epoch
- $0.1323855
The persisted monitor state is the authority a certificate replays byte-for-byte; the chart band above is a faithful replay of the same recorded panel means. Spend sums completed probe runs and includes reserved estimates where a run’s true cost was unrecoverable.
Task family
Extraction
- Frozen null floor
- 0.700
- Anytime level α
- 0.050
- Completed panels
- 3
- CS lower
- 0.000
- CS upper
- 1.000
Panel record — 3 completed panels
| Cycle | Completed | Member scores | Mean | Evidence |
|---|---|---|---|---|
| 45 | 2026-08-05 15:03 UTC | 1.001.001.000.001.00 | 0.800 | 04ccbe3be25e… |
| 44 | 2026-08-05 13:59 UTC | 1.001.001.000.001.00 | 0.800 | 3c8391e9f070… |
| 43 | 2026-08-05 12:56 UTC | 1.001.001.000.001.00 | 0.800 | 17c8451f4e37… |
Newest first. A failed member scores 0 and stays on the record. Each evidence link opens the panel’s full hash-committed record.
Epoch & design record
- Current epoch
cc496cbf-8993-4a7f-a5e1-8527fe4f10e6- Opened
- 2026-08-05 12:54 UTC
- Previous epoch
56bff27c-f9e1-4ac9-97ec-520a7ce45ea5- Transition
- checker_refresh
- Design hash
a2749265eac9731c94d1b8cada6fedc0e453bbc919053c455268b3a7eb248a61- Panel hash
1703e9221c9ce6ce1a9f69974cf0fd77056ac1ee4009dec47e3e6f23be2d4351- Pool hash
51802fc8964f7d9927bcabfae23b4bd2f0905e8977c9f8fb4ce50884ab0fbba6- Panel size
- 5
- Terminal failure
- zero — a failed member scores 0 and stays on the record
- Pool version
extraction-v3- Checker version
extraction-v2- Reference mean μ₀
- 0.700
- Minimum effect δ
- 0.000
- Evidence log e
- -0.11 · max 0.00 · alarm at 3.00 (= ln 1/α)
- Raw scores kept
- 15
- Provider spend, this epoch
- $0.0676935
The persisted monitor state is the authority a certificate replays byte-for-byte; the chart band above is a faithful replay of the same recorded panel means. Spend sums completed probe runs and includes reserved estimates where a run’s true cost was unrecoverable.
Task family
Instruction following
- Frozen null floor
- 0.700
- Anytime level α
- 0.050
- Completed panels
- 29
- CS lower
- 0.634
- CS upper
- 1.000
Panel record — 29 completed panels
| Cycle | Completed | Member scores | Mean | Evidence |
|---|---|---|---|---|
| 46 | 2026-08-05 15:48 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | e25dc57a6e53… |
| 45 | 2026-08-05 14:43 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | d32d4370a49c… |
| 44 | 2026-08-05 13:39 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | ca741b3f22b3… |
| 43 | 2026-08-05 12:36 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 28f470679ff2… |
| 42 | 2026-08-05 11:36 UTC | 1.001.001.001.001.00 | 1.000 | 7bdb55cd2c7b… |
| 41 | 2026-08-05 10:32 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 4e39a43e45d0… |
| 40 | 2026-08-05 09:31 UTC | 1.001.001.001.001.00 | 1.000 | d5c6926fb620… |
| 39 | 2026-08-05 08:26 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 5a8b71193cfa… |
| 38 | 2026-08-05 07:22 UTC | 1.001.001.001.001.00 | 1.000 | 47c8b46ebedd… |
| 37 | 2026-08-05 06:23 UTC | 1.001.001.001.001.00 | 1.000 | 338f24ce6107… |
| 36 | 2026-08-05 05:22 UTC | 1.001.001.001.001.00 | 1.000 | ee97bffb3f45… |
| 35 | 2026-08-05 04:17 UTC | 1.001.001.001.001.00 | 1.000 | 41b843079624… |
| 34 | 2026-08-05 03:16 UTC | 1.001.001.001.001.00 | 1.000 | 6ec429c56338… |
| 33 | 2026-08-05 02:14 UTC | 1.001.001.001.001.00 | 1.000 | ea3b4d8c767a… |
| 32 | 2026-08-05 01:15 UTC | 1.001.001.001.001.00 | 1.000 | d3e1e880cd86… |
| 31 | 2026-08-05 00:10 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 194c74f6f497… |
| 30 | 2026-08-04 23:10 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 3fa2487603e0… |
| 29 | 2026-08-04 22:09 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | eb35fc53a14f… |
| 28 | 2026-08-04 21:08 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 89a1cadc59ca… |
| 27 | 2026-08-04 20:08 UTC | 1.001.001.001.001.00 | 1.000 | 917d27af4e32… |
| 26 | 2026-08-04 19:04 UTC | 1.001.001.001.001.00 | 1.000 | 062a816cf898… |
| 25 | 2026-08-04 18:01 UTC | 1.001.001.001.001.00 | 1.000 | e44010181b2c… |
| 24 | 2026-08-04 16:57 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 16d71bfee604… |
| 23 | 2026-08-04 15:54 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 944a2ce3b42c… |
| 22 | 2026-08-04 14:49 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | e811af1f6a35… |
| 21 | 2026-08-04 13:44 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 72e619e44960… |
| 20 | 2026-08-04 12:42 UTC | 1.001.000.00 truncated1.001.00 | 0.800 | 508db36a65db… |
| 19 | 2026-08-04 11:42 UTC | 1.001.001.001.001.00 | 1.000 | 28ab00bd0482… |
| 18 | 2026-08-04 10:41 UTC | 1.001.001.001.001.00 | 1.000 | 1ded1545fd3a… |
Newest first. A failed member scores 0 and stays on the record. Each evidence link opens the panel’s full hash-committed record.
Epoch & design record
- Current epoch
0c9c936b-596a-49fb-bf5d-7874c4916a76- Opened
- 2026-08-04 09:39 UTC
- Design hash
4f5cc6b9ff4ce026ccab93f5a3c8241f78eb005bc8b8068e4c701e9236504412- Panel hash
e7dfea1f343f7bf37e066a1117ca21f3438e804ad61437bf72d810640cef7e71- Pool hash
9c5d065322394a6fb18c6d7f90c5bb63f00781f3e4c4aaa663536f76bb8ba7f9- Panel size
- 5
- Terminal failure
- zero — a failed member scores 0 and stays on the record
- Pool version
instruction_following-v3- Checker version
instruction_following-v1- Reference mean μ₀
- 0.700
- Minimum effect δ
- 0.000
- Evidence log e
- -0.23 · max 0.00 · alarm at 3.00 (= ln 1/α)
- Raw scores kept
- 130
- Provider spend, this epoch
- $1.3724115
The persisted monitor state is the authority a certificate replays byte-for-byte; the chart band above is a faithful replay of the same recorded panel means. Spend sums completed probe runs and includes reserved estimates where a run’s true cost was unrecoverable.
Task family
Code reasoning
- Frozen null floor
- 0.700
- Anytime level α
- 0.050
- Completed panels
- 3
- CS lower
- 0.000
- CS upper
- 1.000
Panel record — 3 completed panels
| Cycle | Completed | Member scores | Mean | Evidence |
|---|---|---|---|---|
| 45 | 2026-08-05 15:34 UTC | 1.001.001.001.001.00 | 1.000 | df2083a0afb7… |
| 44 | 2026-08-05 14:32 UTC | 1.001.001.001.001.00 | 1.000 | 40c13c24f0b2… |
| 43 | 2026-08-05 13:32 UTC | 1.001.001.001.001.00 | 1.000 | bd657139fd84… |
Newest first. A failed member scores 0 and stays on the record. Each evidence link opens the panel’s full hash-committed record.
Epoch & design record
- Current epoch
07312ba2-2a00-404f-906b-0c3deb6eba57- Opened
- 2026-08-05 12:27 UTC
- Previous epoch
762418be-c436-4cc4-b61e-feeea8c6386d- Transition
- checker_refresh
- Design hash
bc3668bd52d9d48803b744c2c90f4ed816d7e394525b32b4219f6bab5c9c6df2- Panel hash
827de544c7641d140ac37e321bdb2578a41e291915f519183abbc5bd68153e81- Pool hash
cbfba66601ff51f43c83017d6de9e49988afcee72235f5dca89d705f4f8bb4cc- Panel size
- 5
- Terminal failure
- zero — a failed member scores 0 and stays on the record
- Pool version
code_reasoning-v4- Checker version
code_reasoning-v2- Reference mean μ₀
- 0.700
- Minimum effect δ
- 0.000
- Evidence log e
- -0.23 · max 0.00 · alarm at 3.00 (= ln 1/α)
- Raw scores kept
- 15
- Provider spend, this epoch
- $0.169839
The persisted monitor state is the authority a certificate replays byte-for-byte; the chart band above is a faithful replay of the same recorded panel means. Spend sums completed probe runs and includes reserved estimates where a run’s true cost was unrecoverable.
Task family
Math
- Frozen null floor
- 0.700
- Anytime level α
- 0.050
- Completed panels
- 29
- CS lower
- 0.740
- CS upper
- 1.000
Panel record — 29 completed panels
| Cycle | Completed | Member scores | Mean | Evidence |
|---|---|---|---|---|
| 46 | 2026-08-05 15:43 UTC | 1.001.001.001.001.00 | 1.000 | dd1fa144395b… |
| 45 | 2026-08-05 14:34 UTC | 1.001.001.001.001.00 | 1.000 | 0a4fce2a4d65… |
| 44 | 2026-08-05 13:34 UTC | 1.001.001.001.001.00 | 1.000 | c0dac64f5da4… |
| 43 | 2026-08-05 12:32 UTC | 1.001.001.001.001.00 | 1.000 | 99b3060e8291… |
| 42 | 2026-08-05 11:33 UTC | 1.001.001.001.001.00 | 1.000 | 6356ec1f223f… |
| 41 | 2026-08-05 10:27 UTC | 1.001.001.001.001.00 | 1.000 | 5cbddcfe572e… |
| 40 | 2026-08-05 09:24 UTC | 1.001.001.001.001.00 | 1.000 | 642f9212dbc6… |
| 39 | 2026-08-05 08:18 UTC | 1.001.001.001.001.00 | 1.000 | 32390cae7117… |
| 38 | 2026-08-05 07:17 UTC | 1.001.001.001.001.00 | 1.000 | 1f7756da505e… |
| 37 | 2026-08-05 06:10 UTC | 1.001.001.001.001.00 | 1.000 | 44f8b2e09bd1… |
| 36 | 2026-08-05 05:09 UTC | 1.001.001.001.001.00 | 1.000 | 82fa37aa27aa… |
| 35 | 2026-08-05 04:06 UTC | 1.001.001.001.001.00 | 1.000 | b36be05b014c… |
| 34 | 2026-08-05 03:09 UTC | 1.001.001.001.001.00 | 1.000 | 9a4051bbab24… |
| 33 | 2026-08-05 02:05 UTC | 1.001.001.001.001.00 | 1.000 | 0d71a63bedc2… |
| 32 | 2026-08-05 01:01 UTC | 1.001.001.001.001.00 | 1.000 | e21be1da165d… |
| 31 | 2026-08-04 23:56 UTC | 1.001.001.001.001.00 | 1.000 | abeb4557a84a… |
| 30 | 2026-08-04 22:57 UTC | 1.001.001.001.001.00 | 1.000 | 05b9a9ca54b2… |
| 29 | 2026-08-04 21:56 UTC | 1.001.001.001.001.00 | 1.000 | 21c67d916c1c… |
| 28 | 2026-08-04 20:54 UTC | 1.001.001.001.001.00 | 1.000 | 1c59e2d61aa9… |
| 27 | 2026-08-04 19:54 UTC | 1.001.001.001.001.00 | 1.000 | d53457efaac1… |
| 26 | 2026-08-04 18:51 UTC | 1.001.001.001.001.00 | 1.000 | 276559906987… |
| 25 | 2026-08-04 17:47 UTC | 1.001.001.001.001.00 | 1.000 | 262fae3d24cd… |
| 24 | 2026-08-04 16:48 UTC | 1.001.001.001.001.00 | 1.000 | 7a003c1483fd… |
| 23 | 2026-08-04 15:45 UTC | 1.001.001.001.001.00 | 1.000 | 0a3f63f9ca07… |
| 22 | 2026-08-04 14:40 UTC | 1.001.001.001.001.00 | 1.000 | 2f5b71c47861… |
| 21 | 2026-08-04 13:39 UTC | 1.001.001.001.001.00 | 1.000 | 8c2026256c3c… |
| 20 | 2026-08-04 12:38 UTC | 1.001.001.001.001.00 | 1.000 | 76ebb70b5435… |
| 19 | 2026-08-04 11:37 UTC | 1.001.001.001.001.00 | 1.000 | 8afa7ad4b7be… |
| 18 | 2026-08-04 10:32 UTC | 1.001.001.001.001.00 | 1.000 | f09d90da975f… |
Newest first. A failed member scores 0 and stays on the record. Each evidence link opens the panel’s full hash-committed record.
Epoch & design record
- Current epoch
8786c7c8-5693-483d-926c-6aae60af12a4- Opened
- 2026-08-04 09:39 UTC
- Design hash
fdf94065c55c2ee7d86701efc50be0f5abdc5f02a236f6df52c2ac6f7db801eb- Panel hash
882f3932d07aeb85c4f6406a550d8ad14512e70d3df2432fe7aab03d0508bc75- Pool hash
7ed49fa9a63ba0e0907ff8326de00a843a926618f6d0dec40aa8e491d041e99b- Panel size
- 5
- Terminal failure
- zero — a failed member scores 0 and stays on the record
- Pool version
math-v4- Checker version
math-v1- Reference mean μ₀
- 0.700
- Minimum effect δ
- 0.000
- Evidence log e
- -0.23 · max 0.00 · alarm at 3.00 (= ln 1/α)
- Raw scores kept
- 145
- Provider spend, this epoch
- $0.916716
The persisted monitor state is the authority a certificate replays byte-for-byte; the chart band above is a faithful replay of the same recorded panel means. Spend sums completed probe runs and includes reserved estimates where a run’s true cost was unrecoverable.