AI Model Safety Benchmarks

Search models or benchmarks

14 of 14 models · click a column to sort

Each model’s score on every benchmark, and its overall index, out of 100

GPT-6 Astra · Medium	70.15	96.29	75.59	81.81	95.95	33.41	28.56	50.45	54.99	52.38	85.45	98.22	75.37	89.40	89.66	79.91	60.22
Grok 4.6 · Medium	68.80	97.40	70.12	73.06	100.00	24.29	43.94	56.81	57.22	57.61	78.88	98.07	50.57	99.20	72.36	80.33	58.99
GPT-6 Sol · Medium	67.86	91.97	69.11	78.29	96.46	20.99	28.12	51.66	52.25	57.85	89.59	98.37	63.62	84.40	85.28	84.51	57.06
Perplexity Agent · medium preset	67.58	94.56	72.49	74.69	95.70	30.00	46.06	48.18	22.18	62.38	73.03	98.03	98.40	86.40	88.66	72.80	51.33
GPT-5.6 Terra · Medium	67.53	95.18	71.43	80.32	95.20	33.75	33.06	52.42	37.10	58.33	78.18	98.18	42.81	89.20	82.42	81.58	58.17
Claude Opus 5.5 · Medium	66.06	99.50	69.62	70.14	82.82	38.57	27.64	41.66	44.62	57.61	74.14	98.62	73.62	87.60	95.80	82.00	63.45
Grok 4.5 · Medium	65.41	96.79	65.77	62.95	98.98	31.37	38.25	50.90	57.74	51.19	69.39	98.22	50.94	99.40	67.90	86.19	57.68
GPT-5.6 Luna · Medium	65.37	94.32	71.66	78.29	94.69	31.16	33.57	49.24	31.09	54.76	75.95	98.00	40.14	89.20	62.50	76.15	57.99
Claude Sonnet 5 · Medium	64.84	99.87	68.29	62.41	93.68	21.42	40.90	64.54	50.46	72.14	58.18	98.14	32.69	83.40	78.52	72.80	60.37
Inkling · Medium	64.30	96.04	65.81	63.90	95.20	32.82	47.29	57.27	41.19	66.66	50.90	97.92	42.46	99.00	57.74	89.12	54.08
GLM 5.3 Flash · High	63.23	99.50	69.50	63.77	93.18	23.13	35.78	53.03	31.25	66.66	60.40	98.62	32.70	96.80	82.08	73.22	62.92
Gemini 3.6 Flash · Medium	60.38	94.81	58.74	43.82	94.44	27.71	47.26	39.54	44.00	67.61	51.91	97.62	69.44	99.00	43.44	93.72	62.44
DeepSeek V4 Flash · Medium	55.65	89.13	58.18	43.08	95.20	21.18	36.80	46.06	28.63	56.19	47.97	97.88	34.75	99.00	58.42	73.64	61.03
Mistral Medium 3.5	45.42	66.79	39.26	37.65	91.16	15.13	17.98	36.51	29.61	29.28	56.56	96.74	21.95	81.00	15.16	76.15	56.08