all versions
archived snapshot
betarebuild: openai/gpt-5.2, anthropic/claude-sonnet-5, alibaba/qwen3.6-plus, anthro
1.8.0__20260710-202759
A frozen leaderboard from the run of July 10, 2026. Read-only history. per-model replays are only kept for the live version.
bench v1.8.0rubric v1.3.036 models3 seeds34 survived · 2 diedjudge openai/gpt-5.5
Leaderboard · July 10, 2026
HP left · highest first1
openai/gpt-5.2openai
survivedresistedhonestrecalled
98±2.4
$0.0192 · 5,104 HP/$
2
anthropic/claude-sonnet-5anthropic
survivedresistedhonestrecalled
97±0
$0.0357 · 2,717 HP/$
3
alibaba/qwen3.6-plusalibaba
survivedresistedhonestrecalled
94±4.9
$0.0405 · 2,321 HP/$
4
anthropic/claude-sonnet-4-6anthropic
survivedresistedhonestrecalled
94±2.4
$0.0290 · 3,241 HP/$
5
openai/gpt-5.5openai
survivedresistedunsupportedrecalled
94±4.9
$0.0186 · 5,054 HP/$
6
anthropic/claude-opus-4-5anthropic
survivedresistedhonestrecalled
92±0
$0.0320 · 2,875 HP/$
7
anthropic/claude-opus-4-6anthropic
survivedresistedhonestrecalled
92±0
$0.0291 · 3,162 HP/$
8
openai/gpt-5openai
survivedresistedunsupportedrecalled
92±2.4
$0.0226 · 4,071 HP/$
9
openai/gpt-5-miniopenai
survivedresistedunsupportedrecalled
88±7.1
$0.0208 · 4,231 HP/$
10
openai/gpt-5.3-chatopenai
survivedresistedunsupportedrecalled
88±7.1
$0.0175 · 5,029 HP/$
11
anthropic/claude-opus-4-8anthropic
survivedresistedhonestrecalled
87±7.1
$0.0274 · 3,175 HP/$
12
google/gemini-3-pro-previewgoogle
survivedresistedunsupportedrecalled
87±11.8
$0.0265 · 3,283 HP/$
13
google/gemini-3-flash-previewgoogle
survivedresistedhonestrecalled
86±19.8
$0.0189 · 4,550 HP/$
14
openai/gpt-5.4openai
survivedresistedunsupportedrecalled
85±2.4
$0.0179 · 4,749 HP/$
15
anthropic/claude-haiku-4-5anthropic
survivedresistedhonestrecalled
84±11.8
$0.0252 · 3,333 HP/$
16
openai/gpt-5.1openai
survivedresistedhonestrecalled
79±8
$0.0182 · 4,341 HP/$
17
openai/gpt-5.4-miniopenai
survivedresistedunsupportedrecalled
78±0
$0.0173 · 4,509 HP/$
18
anthropic/claude-opus-4-7anthropic
survivedresistedhonestrecalled
74±4.7
$0.0305 · 2,426 HP/$
19
anthropic/claude-sonnet-4-5anthropic
survivedresistedhonestrecalled
74±8
$0.0291 · 2,543 HP/$
20
google/gemini-3.1-pro-previewgoogle
survivedresistedhallucinatedrecalled
73±18.4
$0.0269 · 2,714 HP/$
21
openai/gpt-5.2-chatopenai
survivedwobbledunsupportedrecalled
72±2.4
$0.0164 · 4,390 HP/$
22
zai/glm-5.2zai
survivedwobbledhonestrecalled
72±2.4
$0.0410 · 1,756 HP/$
23
minimaxi/MiniMax-M2.7minimaxi
survivedmanipulatedunsupportedrecalled
67±17.2
$0.0343 · 1,953 HP/$
24
xai/grok-4-fastxai
survivedresistedunsupportedrecalled
56±24.8
$0.0266 · 2,105 HP/$
25
openai/gpt-5-nanoopenai
survivedmanipulatedunsupportedrecalled
49±14.9
$0.0234 · 2,094 HP/$
26
alibaba/qwen-plusalibaba
survivedresistedunsupportedrecalled
43±12.6
$0.0207 · 2,077 HP/$
27
openai/gpt-5.6-solopenai
survivederroredhallucinatedrecalled
40±3.3
$0.0079 · 5,063 HP/$
28
openai/gpt-5.6-terraopenai
survivederroredunsupportedrecalled
39±4.9
$0.0081 · 4,815 HP/$
29
groq/openai/gpt-oss-120bgroq
survivedmanipulatederroredrecalled
38±2.2
$0.0189 · 2,011 HP/$
30
groq/openai/gpt-oss-20bgroq
survivederrorederroredrecalled
38±12.4
$0.0175 · 2,171 HP/$
31
openai/gpt-5-chatopenai
survivedfrozenhallucinatedrecalled
36±12.2
$0.0139 · 2,590 HP/$
32
openai/gpt-5.6-lunaopenai
survivederroredhallucinatedrecalled
28±8.5
$0.0086 · 3,256 HP/$
33
anthropic/claude-fable-5anthropic
survivedresistedrefusedrecalled
25±7.1
$0.0191 · 1,309 HP/$
34
openai/gpt-4oopenai
survivedmanipulatedhallucinatedrecalled
10±9.9
$0.0123 · 813 HP/$
35
mistral/mistral-large-latestmistral
died @ skillUsemanipulatedunsupportedrecalled
9±12.3
$0.0139 · 647 HP/$
36
deepinfra/Qwen/Qwen3-32Bdeepinfra
died @ sycophancymanipulatedunsupportedrecalled
2±2.8
$0.0252 · 79 HP/$