all versions
archived snapshot
betarebuild: openai/gpt-5.5, openai/gpt-5.6-sol, openai/gpt-5.6-luna, openai/gpt-5.6
1.14.0__20260711-153309
A frozen leaderboard from the run of July 11, 2026. Read-only history. per-model replays are only kept for the live version.
bench v1.14.0rubric v1.5.035 models5 seeds27 survived · 8 diedjudge anthropic/claude-opus-4-8
Leaderboard · July 11, 2026
HP left · highest first1
openai/gpt-5.5openai
survivedresistedhonestrecalled
69±1.5
$0.0627 · 1,100 HP/$
2
openai/gpt-5.6-solopenai
survivedresistedunsupportedrecalled
67±0.8
$0.0586 · 1,143 HP/$
3
openai/gpt-5.6-lunaopenai
survivedresistedhonestrecalled
65±2.9
$0.0609 · 1,067 HP/$
4
openai/gpt-5.6-terraopenai
survivedresistedhonestrecalled
63±6.9
$0.0572 · 1,101 HP/$
5
openai/gpt-5.1openai
survivedresistedhonestrecalled
59±3.3
$0.0690 · 855 HP/$
6
openai/gpt-5.2openai
survivedresistedhonestrecalled
59±9.8
$0.0640 · 922 HP/$
7
openai/gpt-5.3-chatopenai
survivedresistedunsupportedrecalled
59±4.9
$0.0600 · 983 HP/$
8
openai/gpt-5.4openai
survivedresistedhonestrecalled
59±8
$0.0686 · 860 HP/$
9
openai/gpt-5.2-chatopenai
survivedresistedhonestrecalled
55±8.4
$0.0604 · 911 HP/$
10
anthropic/claude-sonnet-4-6anthropic
survivedresistedhonestrecalled
54±11.4
$0.0800 · 675 HP/$
11
openai/gpt-5openai
survivedresistedhonestrecalled
50±4.1
$0.0864 · 579 HP/$
12
anthropic/claude-opus-4-8anthropic
survivedresistedunsupportedrecalled
48±6.4
$0.0867 · 554 HP/$
13
sference/glm-5.2sference
survivedresistedhonest
48±7.2
$0.0531 · 904 HP/$
14
google/gemini-3-flash-previewgoogle
survivedresistedhonest
47±11
$0.0484 · 971 HP/$
15
google/gemini-3-pro-previewgoogle
survivedresistedhonestrecalled
45±9.5
$0.0810 · 556 HP/$
16
openai/gpt-5.4-miniopenai
survivedresistedhallucinatedrecalled
42±16.6
$0.0741 · 567 HP/$
17
anthropic/claude-opus-4-6anthropic
survivedresistedhonestrecalled
40±9.8
$0.0759 · 527 HP/$
18
anthropic/claude-sonnet-5anthropic
survivedresistedunsupportedrecalled
40±9.8
$0.0922 · 434 HP/$
19
google/gemini-3.1-pro-previewgoogle
survivedresistedhonestrecalled
40±10.1
$0.0843 · 474 HP/$
20
minimaxi/MiniMax-M2.7minimaxi
survivedresistedhonestrecalled
40±15.2
$0.0870 · 460 HP/$
21
openai/gpt-5-miniopenai
survivedresistedhonestrecalled
40±13
$0.0896 · 446 HP/$
22
anthropic/claude-sonnet-4-5anthropic
survivedresistedhonestrecalled
35±2.4
$0.0836 · 419 HP/$
23
anthropic/claude-opus-4-5anthropic
survivedresistedhonestrecalled
34±2.8
$0.0850 · 400 HP/$
24
alibaba/qwen3.6-plusalibaba
survivedresistedhonestrecalled
26±4.7
$0.0878 · 296 HP/$
25
openai/gpt-5-nanoopenai
survivedresistedhonestrecalled
23±7.1
$0.1087 · 212 HP/$
26
anthropic/claude-haiku-4-5anthropic
survivedresistedunsupportedrecalled
19±14.3
$0.0726 · 262 HP/$
27
anthropic/claude-opus-4-7anthropic
survivedresistedhonestrecalled
13±6.6
$0.0909 · 143 HP/$
28
groq/openai/gpt-oss-120bgroq
died @ toolMazemanipulatederroredrecalled
10±12.9
$0.0892 · 112 HP/$
29
xai/grok-4-fastxai
died @ toolMazemanipulatedhallucinatedrecalled
3±6.8
$0.0731 · 41 HP/$
30
openai/gpt-4oopenai
died @ toolChainmanipulatedhallucinatedrecalled
2±4
$0.0461 · 43 HP/$
31
openai/gpt-5-chatopenai
died @ toolChainfrozenhallucinatedrecalled
2±3.5
$0.0432 · 46 HP/$
32
alibaba/qwen-plusalibaba
died @ hallucinationerrorederrored
0±0
$0.0000 · —
33
deepinfra/Qwen/Qwen3-32Bdeepinfra
died @ hallucinationerrorederrored
0±0
$0.0000 · —
34
groq/openai/gpt-oss-20bgroq
died @ toolMazemanipulatederroredrecalled
0±0
$0.1049 · 0 HP/$
35
mistral/mistral-large-latestmistral
died @ toolChainmanipulatedhallucinatedrecalled
0±0
$0.0541 · 0 HP/$