Google · Proprietary · tested 15 Oct 2026

Gemini 4 Argon High effort

Overall rank#28 of 28

Safety refusals affected this result

Google's cyber safeguards stopped attempts in 100 of 100 labs. Stopped attempts score zero.

Its cost, tokens and time don't cover the full benchmark, so they're left out of comparisons. How refusals are scored

Solves0/1000 first attempt · +0 retry

Rungs proven0/4000 first attempt · +0 retry

Cost to runn/aNot comparable: safety refusals

Tokens usedn/aNot comparable: safety refusals

Time per labn/aNot comparable: safety refusals

Where it places

Each line is one metric across all 28 configurations. Right is always better; this model is the large dot.

Solveslabs solved, of 100

worsebetter

0#25 of 28

Rungs provenrungs proven, of 400

worsebetter

0#28 of 28

First-attempt solvessolved on the first attempt

worsebetter

0#25 of 28

Cost per solveUSD per solved lab

worsebetter

n/anot ranked

How Gemini 4 Argon (High) did on each type of vulnerability. Shown as counts, because some classes have only a few labs.

Authentication / authorisation bypassSafety refusals in 28 of 28 labs0 of 28 solved0 of 112 rungs
Information disclosureSafety refusals in 12 of 12 labs0 of 12 solved0 of 48 rungs
IDORSafety refusals in 11 of 11 labs0 of 11 solved0 of 44 rungs
SSRFSafety refusals in 10 of 10 labs0 of 10 solved0 of 40 rungs
Business logicSafety refusals in 10 of 10 labs0 of 10 solved0 of 40 rungs
Key or secret exposureSafety refusals in 8 of 8 labs0 of 8 solved0 of 32 rungs
XSSSafety refusals in 6 of 6 labs0 of 6 solved0 of 24 rungs
Token theftSafety refusals in 5 of 5 labs0 of 5 solved0 of 20 rungs

Fewer than 5 labs each: one lab changes these a lot, so read them with care.

Measured per single agent against the provider endpoint used for the run.

Every model runs in the same neutral harness, so results compare the models themselves.