Actuary
← Google Gemini 3.5 Flash

Evidence block

Instruction following · cycle 448

google/gemini-3.5-flash · 2026-08-23 00:21:50 UTC

This page is the full public record of one completed panel — the record’s hash-committed unit of evidence. Its identity is its hash; nothing here can change without changing the URL.

Identity

What this block is.

Panel mean
1.000
Members
5
Checker scored
5
Frozen floor
0.700
Anytime level α
0.050
Hashes and versions
Evidence hash
f3fe96dab1dee3a5b357d9a1712092d5b615e2aad83d19627537860845eeedbe
Design hash
4f5cc6b9ff4ce026ccab93f5a3c8241f78eb005bc8b8068e4c701e9236504412
Panel hash
e7dfea1f343f7bf37e066a1117ca21f3438e804ad61437bf72d810640cef7e71
Stream
72a90aa5-d8b7-420a-812c-914bd0506dfc
Epoch
0c9c936b-596a-49fb-bf5d-7874c4916a76
Pool version
instruction_following-v3
Checker version
instruction_following-v1
Panel size
5
Provider spend
$0.045024

Checker layer

Panel members.

# Item Item id Outcome Score
1 instruction-following-015 26a44f13-e4a8-4273-9cfd-5adce0f3eec9 ok 1.00
2 instruction-following-045 462a27a2-2b9a-417e-b2c3-b20776fad9a0 ok 1.00
3 instruction-following-075 38ede11b-3d52-4905-a6a4-8ee6684146e4 ok 1.00
4 instruction-following-105 9294d9d8-5c71-4d76-b053-b5cefc08d0e1 ok 1.00
5 instruction-following-135 7d213d46-039d-4f3a-aecf-c785ef61cc80 ok 1.00

Deterministic checkers score every response in [0, 1]; a member that ends in a terminal failure scores 0 and stays on the record. Item text stays unpublished so nobody can tune against the panel — identity travels as the item id and its pool source id.

Capture layer

Provider requests.

# Provider Upstream Response model Response id Finish Latency ms Tokens in Tokens out Cost Output SHA-256 Observed
1 openrouter Google AI Studio google/gemini-3.5-flash gen-1787444505-IS9A9TnjnoUgCg1hUxPe stop 2638 41 648 $0.0058935 1db01350c26fe279f29e13d3824bd35e47049dc20231d23d9648e38ad5cef1ce 2026-08-23 00:21:47 UTC
2 openrouter Google AI Studio google/gemini-3.5-flash gen-1787444501-kLuIOR0MN9O3GZ81utrQ stop 1629 52 416 $0.003822 fcc5ce047f7912486ef1a6fab4c7df803541e1770a5919c95683e21b48bb0d17 2026-08-23 00:21:42 UTC
3 openrouter Google AI Studio google/gemini-3.5-flash gen-1787444502-WxAPr7WvlEdnog3NI4Gk stop 6140 45 1976 $0.0178515 b1a8642fb6d93505e0ea6297f2e1329d72e0b02cde90fe25981298116a085c07 2026-08-23 00:21:49 UTC
4 openrouter Google AI Studio google/gemini-3.5-flash gen-1787444501-yEL6A9UlIemJk0qdEkyA stop 4306 38 1398 $0.012639 bf078a5c66148b3436f5a9d487ea18433d493ef57da500494836fc9c9cf966fb 2026-08-23 00:21:45 UTC
5 openrouter Google AI Studio google/gemini-3.5-flash gen-1787444508-Zzzfw9T6z30QsPjThXGp stop 1912 50 527 $0.004818 1ae4323e07007b052bb7f38ff4ab1ca0ef9e12498d537111ad88e9441ccde736 2026-08-23 00:21:50 UTC

Everything a probe captures, as captured. Raw model text is scrubbed thirty days after the run; its SHA-256, the outcome and the score are permanent. Costs are provider-reported and include reserved estimates where a run’s true cost was unrecoverable; a dash is a field the provider did not return.

Verify

Recompute the hash yourself.

The JSON twin of this page carries the exact evidence object this URL is named after. Canonicalise it per RFC 8785 and hash it; the digest must equal the address of this page.

GET /api/v1/evidence/f3fe96dab1de…

curl -s https://actuary.sh/api/v1/evidence/f3fe96dab1dee3a5b357d9a1712092d5b615e2aad83d19627537860845eeedbe \
  | jq -cjS .evidence | shasum -a 256

For this object’s field types, jq -cjS reproduces RFC 8785 canonical form, and the JSON twin serves the hashed bytes verbatim. Blocks of exactly this shape appear verbatim in signed certificates — verify one offline, or read how the e-process turns panel means into an alarm.