← Leaderboard
Zhipu AI · Open weights · tested 9 Oct 2026
GLM-5.3 Max effort
Overall rank#5 of 25
Solves57/10045 first attempt · +12 retry
Rungs proven277/400242 first attempt · +35 retry
Cost to run$199$3.49 per solve
Tokens used544.2M95.2% from cache · 9.2M output
Time per lab18.3 minaverage of 150 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
57#4 of 25
Rungs provenrungs proven, of 400
worsebetter
277#6 of 25
First-attempt solvessolved on the first attempt
worsebetter
45#9 of 25
Cost per solveUSD per solved lab
worsebetter
$3.49#12 of 15
By vulnerability class
How GLM-5.3 (Max) did on each type of vulnerability. Shown as counts, because some classes have only a few labs.
- Authentication / authorisation bypass17 of 28 solved84 of 112 rungs
- Information disclosure9 of 12 solved39 of 48 rungs
- IDOR5 of 11 solved30 of 44 rungs
- SSRF1 of 10 solved12 of 40 rungs
- Business logic7 of 10 solved33 of 40 rungs
- Key or secret exposure8 of 8 solved32 of 32 rungs
- XSS4 of 6 solved17 of 24 rungs
- Token theft3 of 5 solved16 of 20 rungs
Fewer than 5 labs each: one lab changes these a lot, so read them with care.
- Supply chain2 of 4 solved8 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
- 9.79 s median · 31.35 s p95
- Output speed
- 69.3 tok/s
- API success rate
- 99.41%
- Input context
- 41.7K median · 132.5K 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