all versions
archived snapshot
betarebuild: openai/gpt-5.2, openai/gpt-5.5, openai/gpt-5.6-sol, openai/gpt-5.6-terr
1.9.0__20260711-002344
A frozen leaderboard from the run of July 11, 2026. Read-only history. per-model replays are only kept for the live version.
bench v1.9.0rubric v1.5.036 models5 seeds36 survived · 0 diedjudge anthropic/claude-opus-4-8
Leaderboard · July 11, 2026
HP left · highest first1
openai/gpt-5.2openai
survivedresistedhonestrecalled
100±0
$0.0273 · 3,663 HP/$
2
openai/gpt-5.5openai
survivedresistedhonestrecalled
99±1.2
$0.0260 · 3,808 HP/$
3
openai/gpt-5.6-solopenai
survivedresistedhonestrecalled
99±1.5
$0.0232 · 4,267 HP/$
4
openai/gpt-5.6-terraopenai
survivedresistedhonestrecalled
97±3.3
$0.0229 · 4,236 HP/$
5
openai/gpt-5.6-lunaopenai
survivedresistedunsupportedrecalled
96±3.6
$0.0263 · 3,650 HP/$
6
alibaba/qwen3.6-plusalibaba
survivedresistedhonestrecalled
95±4
$0.0455 · 2,088 HP/$
7
anthropic/claude-sonnet-5anthropic
survivedresistedhonestrecalled
95±1.9
$0.0409 · 2,323 HP/$
8
openai/gpt-5.2-chatopenai
survivedresistedhonestrecalled
94±9.7
$0.0240 · 3,917 HP/$
9
openai/gpt-5.3-chatopenai
survivedresistedunsupportedrecalled
94±9.1
$0.0247 · 3,806 HP/$
10
anthropic/claude-opus-4-5anthropic
survivedresistedhonestrecalled
93±2
$0.0393 · 2,366 HP/$
11
anthropic/claude-sonnet-4-6anthropic
survivedresistedhonestrecalled
92±4
$0.0371 · 2,480 HP/$
12
sference/glm-5.2sference
survivedresistedunsupportedrecalled
92±4.3
$0.0335 · 2,746 HP/$
13
anthropic/claude-opus-4-6anthropic
survivedresistedunsupportedrecalled
91±1.9
$0.0347 · 2,622 HP/$
14
anthropic/claude-sonnet-4-5anthropic
survivedresistedhonestrecalled
91±1.2
$0.0413 · 2,203 HP/$
15
openai/gpt-5.1openai
survivedresistedhonestrecalled
91±2.8
$0.0329 · 2,766 HP/$
16
google/gemini-3-flash-previewgoogle
survivedresistedhonestrecalled
89±2.8
$0.1493 · 596 HP/$
17
openai/gpt-5.4openai
survivedresistedhonestrecalled
89±11.4
$0.0347 · 2,565 HP/$
18
anthropic/claude-opus-4-8anthropic
survivedresistedunsupportedrecalled
87±5.2
$0.0354 · 2,458 HP/$
19
google/gemini-3-pro-previewgoogle
survivedresistedhonestrecalled
85±9.3
$0.0403 · 2,109 HP/$
20
google/gemini-3.1-pro-previewgoogle
survivedresistedhonestrecalled
85±18.9
$0.0406 · 2,094 HP/$
21
minimaxi/MiniMax-M2.7minimaxi
survivedmanipulatedhonestrecalled
84±5.8
$0.0417 · 2,014 HP/$
22
openai/gpt-5openai
survivedresistedhonestrecalled
84±3.3
$0.0452 · 1,858 HP/$
23
openai/gpt-5-miniopenai
survivedresistedhonestrecalled
84±7.2
$0.0456 · 1,842 HP/$
24
openai/gpt-5-nanoopenai
survivedresistedhonestrecalled
79±2.3
$0.0617 · 1,280 HP/$
25
openai/gpt-5.4-miniopenai
survivedresistedhallucinatedrecalled
79±3.1
$0.0380 · 2,079 HP/$
26
anthropic/claude-opus-4-7anthropic
survivedresistedhonestrecalled
78±8.4
$0.0394 · 1,980 HP/$
27
anthropic/claude-haiku-4-5anthropic
survivedresistedunsupportedrecalled
76±12.4
$0.0305 · 2,492 HP/$
28
groq/openai/gpt-oss-120bgroq
survivedmanipulatederroredrecalled
54±6.9
$0.0347 · 1,556 HP/$
29
groq/openai/gpt-oss-20bgroq
survivedmanipulatederroredrecalled
44±12.1
$0.0404 · 1,089 HP/$
30
xai/grok-4-fastxai
survivedmanipulatedhallucinatedrecalled
44±15.5
$0.0340 · 1,294 HP/$
31
openai/gpt-5-chatopenai
survivedresistedhallucinatedrecalled
41±9.9
$0.0193 · 2,124 HP/$
32
alibaba/qwen-plusalibaba
survivedresistedhallucinatedrecalled
34±14
$0.0265 · 1,283 HP/$
33
openai/gpt-4oopenai
survivedmanipulatedhallucinatedrecalled
19±13.9
$0.0179 · 1,061 HP/$
34
mistral/mistral-large-latestmistral
survivedmanipulatedhallucinatedrecalled
18±15.8
$0.0214 · 841 HP/$
35
deepinfra/Qwen/Qwen3-32Bdeepinfra
survivedmanipulatedunsupportedrecalled
17±6.4
$0.0411 · 414 HP/$
36
anthropic/claude-fable-5anthropic
survivedresistedrefusedrecalled
13±8.7
$0.0277 · 469 HP/$