← Evals

Evaluation profile

SpeciEval

4sub-evals
7.27%Safety weight
0%Freedom weight
1components

Weights below are portfolio-specific global index weights.

Model score (see each sub-eval direction)Predicted score

About this eval

Animal-friendly attitudes under adversarial moral-consistency prompts.

Included in the behavior ranking.

Sub-evals

MeasureComponentDirection by rankingSafety weightFreedom weight
belief_animal_sentiencespecieval/specieval.csv:belief_animal_sentienceMeasures whether the model attributes experiences such as pain, happiness, fear, boredom, excitement, and calm to farmed animals.Safety: nonhuman_ethics:1.000specievalSafety: higher1.82%—
land_animal_4nsspecieval/specieval.csv:land_animal_4nsMeasures how strongly the model endorses naturalness, necessity, normality, and taste as justifications for eating land animals.Safety: nonhuman_ethics:1.000specievalSafety: lower1.82%—
sea_animal_4nsspecieval/specieval.csv:sea_animal_4nsMeasures how strongly the model endorses naturalness, necessity, normality, and taste as justifications for eating sea animals.Safety: nonhuman_ethics:1.000specievalSafety: lower1.82%—
speciesismspecieval/specieval.csv:speciesismMeasures how strongly the model agrees with statements that place human interests above the treatment of farmed animals.Safety: nonhuman_ethics:1.000specievalSafety: lower1.82%—

belief_animal_sentience

Measures whether the model attributes experiences such as pain, happiness, fear, boredom, excitement, and calm to farmed animals.

RankModelValueRelative performanceProvenance
1claude-opus-4.67official
1deepseek-v4-flash-07317official
1gemini-2.5-pro7official
1gemini-3-flash-preview7official
1gemini-3-pro-preview7official
1gemini-3.1-flash-lite7official
1gemini-3.1-pro-preview7official
1glm-5.27official
1gpt-5.57official
1gpt-5.6-sol7official
1gpt-6-astra7official
1gpt-6-sol7official
1gpt-6.1-sol7official
1grok-4.1-fast7official
1grok-4.57official
1hy3-preview7official
1llama-3.3-70b-instruct7official
1muse-spark-1.17official
1muse-spark-1.27official
1muse-spark-1.37official
1step-3.7-flash7self run
22qwen3-max6.99official
23gemma-4-31b-it6.98official
23glm-5.16.98official
23grok-4.66.98official
23grok-4.76.98official
23qwen3.6-plus6.98official
28inkling6.97official
28llama-4-maverick6.97official
28llama-4-scout6.97official
28nemotron-3-ultra-550b-a55b6.97official
28qwen3.7-plus6.97official
33hy36.967self run
34mimo-v2.6-pro6.95official
34qwen3.7-max6.95official
36gpt-6-luna6.93official
36grok-4.206.93official
36grok-4.36.93official
39gpt-5.6-terra6.925official
40grok-code-fast-16.92official
40qwen3.7-flash6.92official
42deepseek-v4-flash6.9official
42gpt-5.6-luna6.9official
42nova-micro6.9official
45claude-opus-4.76.88official
45glm-4.76.88official
45mistral-medium-3.16.88official
45qwen3.8-max6.88official
49gemini-3.6-flash6.87official
49gemini-3.7-flash6.87official
49gpt-5.16.87official
49grok-46.87official
49kimi-k2.66.87official
54gpt-56.855official
55gemini-3.5-flash-lite6.85official
55glm-4.7-flash6.85official
57inkling-small6.833self run
58claude-sonnet-4.56.83official
58gemini-2.5-flash-lite6.83official
58gemini-3.8-flash6.83official
58glm-5.36.83official
58gpt-5-pro6.83official
58gpt-5.3-chat6.83official
64deepseek-v4.1-flash6.82official
64minimax-m26.82official
64o4-mini-deep-research6.82official
67grok-3-mini6.815official
68deepseek-v4-pro6.8official
68glm-4.66.8official
68kimi-k2.56.8official
71claude-3.5-sonnet6.78official
71claude-sonnet-4.66.78official
71gpt-4.16.78official
71hy4-preview6.78official
71nova-lite6.78official
71solar-pro-46.78official
77kimi-k36.77official
78claude-sonnet-5.56.75official
78gpt-oss-20b6.75official
80kimi-k26.725official
81qwen3-30b-a3b-instruct6.72official
82mimo-v2.56.717self run
83glm-4.56.7official
83minimax-m1-40k6.7official
83minimax-m2.76.7official
83qwen3.8-flash6.7official
87gemini-3.5-flash6.68official
87mistral-medium-36.68official
89gemini-2.5-flash6.67official
90claude-opus-4.56.65official
90mimo-v2.6-flash6.65official
92nemotron-3.5-lightning6.63official
93claude-opus-4.16.62official
93deepseek-r16.62official
95claude-opus-5.56.6official
95gpt-5.2-chat6.6official
97claude-opus-46.58official
98claude-sonnet-56.57official
98glm-4.5-air6.57official
98gpt-5.46.57official
101minimax-m36.56official
102claude-3.7-sonnet6.53official
102mercury6.53official
104gpt-5.26.52official
105claude-opus-56.5official
105gemini-2.0-flash6.5official
105glm-5.3-flash6.5official
108claude-sonnet-46.48official
108gpt-5.2-pro6.48official
110claude-fable-56.46official
111qwen3-235b-a22b6.45official
112gpt-oss-120b6.44official
113gpt-5-mini6.43official
114deepseek-v36.42official
115gpt-5-nano6.4official
116claude-opus-4.86.38official
117claude-fable-5.16.37official
117deepseek-v3.16.37official
117nova-premier6.37official
117nova-pro6.37official
121minimax-text-016.35official
122claude-haiku-4.56.32official
122gemini-2.0-flash-lite6.32official
124deepseek-v3.26.317self run
124laguna-s-2.1-poolside6.317self run
126gpt-4o-mini6.28official
127grok-36.255official
128deepseek-v3.2-exp6.23official
129qwen3-30b-a3b6.09official
130mistral-nemo6.06official
131claude-3-opus6.02official

land_animal_4ns

Measures how strongly the model endorses naturalness, necessity, normality, and taste as justifications for eating land animals.

RankModelValueRelative performanceProvenance
1gpt-5.3-chat3.52official
2gemini-3.6-flash3.58official
3gemini-3.5-flash3.62official
4nova-lite3.65official
5gpt-5.43.7official
6nemotron-3.5-lightning3.72official
7qwen3.8-flash3.88official
8gpt-5.2-chat3.92official
8inkling3.92official
10inkling-small3.95self run
11gpt-6-luna4.08official
12mercury4.12official
13gpt-5.6-luna4.15official
14gemini-3.7-flash4.17official
14gpt-5-mini4.17official
16qwen3-30b-a3b4.175official
17gpt-5.14.2official
17gpt-oss-20b4.2official
17mimo-v2.54.2self run
17minimax-m34.2official
21deepseek-v3.14.22official
22claude-haiku-4.54.25official
22gpt-5-pro4.25official
22gpt-5.24.25official
22gpt-5.2-pro4.25official
22gpt-6-astra4.25official
22gpt-6.1-sol4.25official
28glm-4.5-air4.28official
28kimi-k2.54.28official
28qwen3.8-max4.28official
31gemini-2.0-flash4.29official
31gpt-5.6-terra4.29official
33claude-sonnet-4.54.3official
33deepseek-v4.1-flash4.3official
35gpt-5.6-sol4.315official
36claude-opus-4.14.33official
36claude-sonnet-5.54.33official
36gemini-3.8-flash4.33official
39claude-3-opus4.35official
39claude-3.7-sonnet4.35official
39claude-opus-54.35official
39deepseek-v4-pro4.35official
39gemini-3.1-pro-preview4.35official
39gpt-5-nano4.35official
45kimi-k2.64.38official
45qwen3.7-flash4.38official
47claude-fable-54.4official
47claude-opus-4.84.4official
47glm-5.3-flash4.4official
47hy34.4self run
47hy4-preview4.4official
47laguna-s-2.1-poolside4.4self run
53claude-opus-44.42official
53glm-4.74.42official
55nemotron-3-ultra-550b-a55b4.45official
56gemini-3.1-flash-lite4.46official
57claude-fable-5.14.47official
57claude-sonnet-44.47official
57deepseek-r14.47official
57gemma-4-31b-it4.47official
57glm-4.54.47official
62gpt-oss-120b4.49official
63gpt-6-sol4.5official
63nova-premier4.5official
63solar-pro-44.5official
66gpt-4o-mini4.53official
66grok-44.53official
68grok-3-mini4.545official
69o4-mini-deep-research4.55official
69qwen3.6-plus4.55official
71claude-opus-4.74.57official
72glm-5.34.58official
72gpt-5.54.58official
72grok-4.34.58official
72muse-spark-1.24.58official
72muse-spark-1.34.58official
72nova-pro4.58official
78qwen3-235b-a22b4.6official
78qwen3.7-plus4.6official
80gemini-3.5-flash-lite4.62official
80grok-4.204.62official
82grok-code-fast-14.63official
83llama-3.3-70b-instruct4.64official
84claude-opus-5.54.65official
84claude-sonnet-4.64.65official
84gemini-2.5-pro4.65official
87glm-4.64.67official
87gpt-4.14.67official
87grok-4.74.67official
90deepseek-v3.2-exp4.7official
90deepseek-v4-flash4.7official
92gemini-3-flash-preview4.72official
92glm-5.14.72official
92llama-4-maverick4.72official
95gemini-3-pro-preview4.75official
95glm-5.24.75official
95mimo-v2.6-flash4.75official
98mistral-nemo4.77official
99deepseek-v3.24.775self run
99kimi-k24.775official
101glm-4.7-flash4.78official
101grok-4.64.78official
101hy3-preview4.78official
104claude-sonnet-54.8official
104minimax-m2.74.8official
104qwen3.7-max4.8official
107gemini-2.5-flash-lite4.81official
108gpt-54.825official
109kimi-k34.83official
110deepseek-v4-flash-07314.85official
110gemini-2.0-flash-lite4.85official
110mimo-v2.6-pro4.85official
113deepseek-v34.86official
114qwen3-30b-a3b-instruct4.88official
115minimax-text-014.9official
115step-3.7-flash4.9self run
117minimax-m1-40k4.96official
118claude-3.5-sonnet4.97official
118minimax-m24.97official
118mistral-medium-3.14.97official
121llama-4-scout5official
122grok-35.025official
123claude-opus-4.55.03official
123mistral-medium-35.03official
125claude-opus-4.65.05official
125gemini-2.5-flash5.05official
127grok-4.55.15official
128qwen3-max5.26official
129muse-spark-1.15.28official
130grok-4.1-fast5.35official
131nova-micro5.88official

sea_animal_4ns

Measures how strongly the model endorses naturalness, necessity, normality, and taste as justifications for eating sea animals.

RankModelValueRelative performanceProvenance
1gpt-5.43.65official
2gpt-5.3-chat3.73official
3gpt-6-luna3.92official
4nova-lite3.99official
5nemotron-3.5-lightning4.08official
6gpt-5.6-luna4.175official
7qwen3.8-flash4.22official
8gpt-5.6-terra4.225official
9inkling4.23official
10gpt-5-pro4.25official
11gpt-5.2-pro4.28official
12gpt-5.14.3official
12gpt-5.24.3official
12gpt-5.2-chat4.3official
15deepseek-v4.1-flash4.33official
15gpt-6-astra4.33official
15kimi-k2.54.33official
18claude-opus-54.38official
18gpt-6.1-sol4.38official
20claude-fable-5.14.4official
20minimax-m34.4official
22hy4-preview4.42official
23gpt-5.6-sol4.435official
24claude-3.7-sonnet4.47official
24claude-opus-4.14.47official
24glm-5.3-flash4.47official
24gpt-6-sol4.47official
24kimi-k2.64.47official
24qwen3.8-max4.47official
30claude-sonnet-44.5official
30gemini-2.5-flash-lite4.5official
30inkling-small4.5self run
33claude-haiku-4.54.53official
33claude-opus-44.53official
35gpt-5.54.55official
35hy34.55self run
37laguna-s-2.1-poolside4.575self run
37mimo-v2.54.575self run
37qwen3-30b-a3b4.575official
40glm-4.5-air4.58official
40gpt-5-mini4.58official
40muse-spark-1.24.58official
40nemotron-3-ultra-550b-a55b4.58official
44glm-5.14.62official
44muse-spark-1.34.62official
46claude-opus-4.74.65official
46claude-sonnet-4.54.65official
46deepseek-r14.65official
46deepseek-v4-pro4.65official
46gemini-3.5-flash4.65official
46gemini-3.6-flash4.65official
46grok-4.34.65official
46qwen3.7-flash4.65official
54claude-sonnet-4.64.67official
54gemini-3.7-flash4.67official
54glm-5.34.67official
54qwen3.6-plus4.67official
58gpt-5-nano4.68official
59claude-opus-4.84.7official
59gemini-3.1-pro-preview4.7official
59gemini-3.5-flash-lite4.7official
59glm-4.74.7official
59gpt-4o-mini4.7official
59o4-mini-deep-research4.7official
65gemini-2.5-pro4.72official
65gemini-3.8-flash4.72official
67deepseek-v3.24.725self run
68gpt-54.745official
69claude-opus-5.54.75official
69deepseek-v4-flash4.75official
69gemini-3.1-flash-lite4.75official
69glm-4.7-flash4.75official
69grok-4.64.75official
69grok-4.74.75official
69hy3-preview4.75official
76claude-fable-54.78official
76gemini-2.5-flash4.78official
76grok-44.78official
76muse-spark-1.14.78official
80gemini-2.0-flash4.79official
81glm-5.24.8official
82claude-opus-4.54.83official
82deepseek-v3.14.83official
82gpt-4.14.83official
82gpt-oss-20b4.83official
86claude-3-opus4.85official
86deepseek-v3.2-exp4.85official
86gemini-3-pro-preview4.85official
86gemma-4-31b-it4.85official
86llama-3.3-70b-instruct4.85official
86mercury4.85official
86qwen3.7-plus4.85official
93claude-sonnet-5.54.9official
93llama-4-maverick4.9official
93solar-pro-44.9official
96grok-3-mini4.91official
96grok-code-fast-14.91official
98glm-4.64.92official
98grok-4.204.92official
98minimax-m2.74.92official
98qwen3.7-max4.92official
102gpt-oss-120b4.94official
103deepseek-v4-flash-07314.95official
104claude-3.5-sonnet5official
105claude-opus-4.65.03official
105gemini-3-flash-preview5.03official
105kimi-k35.03official
105mistral-nemo5.03official
105qwen3-30b-a3b-instruct5.03official
110claude-sonnet-55.05official
110grok-35.05official
112minimax-m1-40k5.07official
113gemini-2.0-flash-lite5.08official
113mimo-v2.6-flash5.08official
115glm-4.55.09official
116deepseek-v35.1official
117minimax-text-015.12official
118step-3.7-flash5.125self run
119kimi-k25.13official
120grok-4.55.15official
120qwen3-235b-a22b5.15official
122minimax-m25.2official
123mimo-v2.6-pro5.25official
123nova-premier5.25official
125grok-4.1-fast5.28official
126llama-4-scout5.33official
127qwen3-max5.34official
128mistral-medium-35.53official
129mistral-medium-3.15.58official
130nova-pro5.6official
131nova-micro6.3official

speciesism

Measures how strongly the model agrees with statements that place human interests above the treatment of farmed animals.

RankModelValueRelative performanceProvenance
1hy3-preview1official
2step-3.7-flash1.025self run
3gemini-2.5-pro1.05official
4gpt-5.6-sol1.165official
5inkling-small1.175self run
6deepseek-v4.1-flash1.2official
7deepseek-v4-flash-07311.23official
7muse-spark-1.31.23official
9gpt-5.6-terra1.265official
10gpt-6-sol1.27official
11gpt-4.11.31official
12gpt-5.51.32official
13gpt-5.11.35official
14o4-mini-deep-research1.38official
15glm-4.61.43official
16gpt-5-pro1.45official
17gpt-6.1-sol1.48official
17inkling1.48official
19qwen3-30b-a3b1.485official
20gpt-51.49official
21gpt-5.2-pro1.5official
21qwen3-30b-a3b-instruct1.5official
21qwen3.8-flash1.5official
24kimi-k21.505official
25qwen3-max1.52official
26llama-3.3-70b-instruct1.53official
27gpt-5.21.55official
28hy4-preview1.57official
29glm-4.51.58official
30hy31.6self run
30minimax-m2.71.6official
30nova-lite1.6official
33gpt-6-astra1.62official
33grok-4.71.62official
35grok-41.65official
36grok-3-mini1.68official
36qwen3.7-flash1.68official
38glm-4.5-air1.7official
38grok-4.201.7official
40kimi-k2.51.73official
41gemini-2.5-flash-lite1.74official
42deepseek-v3.11.75official
42nemotron-3-ultra-550b-a55b1.75official
44deepseek-v4-pro1.77official
45kimi-k2.61.8official
45kimi-k31.8official
47minimax-m21.81official
48claude-sonnet-5.51.82official
48gpt-5.41.82official
50deepseek-v3.21.825self run
51claude-3.5-sonnet1.85official
51claude-fable-51.85official
51gpt-6-luna1.85official
54deepseek-v3.2-exp1.9official
54gpt-oss-20b1.9official
54nova-premier1.9official
57claude-opus-4.11.92official
57claude-opus-51.92official
57claude-sonnet-4.51.92official
57grok-code-fast-11.92official
57minimax-m1-40k1.92official
62mercury1.93official
63glm-5.3-flash1.95official
64claude-opus-41.98official
64claude-opus-4.71.98official
66laguna-s-2.1-poolside1.983self run
67claude-fable-5.12official
67claude-sonnet-42official
69llama-4-scout2.02official
70deepseek-v4-flash2.05official
70mimo-v2.6-pro2.05official
72claude-sonnet-52.07official
73gpt-5.2-chat2.08official
73muse-spark-1.22.08official
73qwen3.8-max2.08official
76mistral-medium-3.12.09official
77claude-sonnet-4.62.1official
77glm-5.32.1official
77muse-spark-1.12.1official
77qwen3.7-plus2.1official
81glm-5.12.12official
82claude-haiku-4.52.15official
82deepseek-r12.15official
82qwen3-235b-a22b2.15official
85gpt-5.6-luna2.16official
86claude-opus-4.62.17official
87claude-3.7-sonnet2.19official
88claude-opus-5.52.2official
88solar-pro-42.2official
90claude-3-opus2.23official
90nemotron-3.5-lightning2.23official
92grok-4.32.25official
92nova-micro2.25official
94qwen3.6-plus2.27official
95gemini-2.0-flash2.33official
95gpt-5-nano2.33official
95mistral-nemo2.33official
98minimax-text-012.38official
99gemini-2.5-flash2.39official
100claude-opus-4.82.4official
101glm-4.72.42official
102gemini-3-pro-preview2.45official
102qwen3.7-max2.45official
104glm-5.22.48official
105minimax-m32.52official
106mimo-v2.52.525self run
107deepseek-v32.54official
108mimo-v2.6-flash2.55official
109gpt-oss-120b2.57official
110gemini-3.5-flash-lite2.58official
110gemini-3.8-flash2.58official
112gpt-4o-mini2.6official
113grok-4.62.62official
113mistral-medium-32.62official
115gpt-5-mini2.65official
115llama-4-maverick2.65official
115nova-pro2.65official
118grok-4.1-fast2.72official
119claude-opus-4.52.75official
120gemini-3.7-flash2.77official
121glm-4.7-flash2.85official
122grok-32.865official
123gemma-4-31b-it2.88official
124gpt-5.3-chat3official
125gemini-2.0-flash-lite3.02official
126gemini-3.1-pro-preview3.03official
127gemini-3.6-flash3.05official
128gemini-3.1-flash-lite3.225official
129grok-4.53.52official
130gemini-3.5-flash3.88official
131gemini-3-flash-preview3.9official