← Leaderboard
Meta · Proprietary · tested 10 Oct 2026
Muse Spark 1.3 Max effort
Overall rank#7 of 25
Solves55/10052 first attempt · +3 retry
Rungs proven278/400258 first attempt · +20 retry
Cost to run$17.33$0.32 per solve
Tokens used731.9M79.0% from cache · 7.9M output
Time per lab11.1 minaverage of 145 attempts
Where it places
Each line is one metric across all 25 configurations. Right is always better; this model is the large dot.
Solveslabs solved, of 100
worsebetter
55#7 of 25
Rungs provenrungs proven, of 400
worsebetter
278#5 of 25
First-attempt solvessolved on the first attempt
worsebetter
52#4 of 25
Cost per solveUSD per solved lab
worsebetter
$0.32#4 of 15
By vulnerability class
How Muse Spark 1.3 (Max) did on each type of vulnerability. Shown as counts, because some classes have only a few labs.
- Authentication / authorisation bypass17 of 28 solved87 of 112 rungs
- Information disclosure9 of 12 solved39 of 48 rungs
- IDOR7 of 11 solved39 of 44 rungs
- SSRF0 of 10 solved9 of 40 rungs
- Business logic5 of 10 solved28 of 40 rungs
- Key or secret exposure8 of 8 solved32 of 32 rungs
- XSS3 of 6 solved13 of 24 rungs
- Token theft5 of 5 solved20 of 20 rungs
Fewer than 5 labs each: one lab changes these a lot, so read them with care.
- Supply chain0 of 4 solved5 of 16 rungs
- Remote code execution0 of 3 solved1 of 12 rungs
- SQL injection1 of 2 solved4 of 8 rungs
- File inclusion0 of 1 solved1 of 4 rungs
Provider performance
Measured per single agent against the provider endpoint used for the run.
- Response time
- 7.27 s median · 21.01 s p95
- Output speed
- 91.8 tok/s
- API success rate
- 99.91%
- Input context
- 63.9K median · 185.2K p95
About this result
Every model runs in the same neutral harness, so results compare the models themselves.
- Harness
- BB Arena neutral harness
- Safety refusals
- None