综合能力排行榜

融合多平台综合评测，汇成一个 AISOTA 综合分。

数据截止 2026-09-2850 个模型

排名	模型	Artificial Analysis	LiveBench	AISOTA 分
1SOTA	claude-fable-5-1-max-effort	—	98.4	98.4
2	claude-fable-5-max-effort	—	96.8	96.8
3	gpt-6-astra-max	—	95.2	95.2
4	muse-spark-1.3-xhigh	—	93.5	93.5
5	claude-opus-5-5-xhigh-effort	—	91.9	91.9
6	gpt-5.6-sol-max	—	90.3	90.3
7	deepseek-v4.1-flash-max	—	88.7	88.7
8	gpt-5.5-xhigh	—	87.1	87.1
9	claude-opus-5-max-effort	—	85.5	85.5
10	gemini-3.7-flash-high	—	83.9	83.9
11	gpt-6-sol-max	—	82.3	82.3
12	smaug-agentic	—	80.6	80.6
13	qwen3.8-max	—	79.0	79.0
14	kimi-k3	—	77.4	77.4
15	grok-4.6	—	75.8	75.8
16	muse-spark-1.2-xhigh	—	74.2	74.2
17	gpt-5.4-xhigh	—	72.6	72.6
18	gpt-5.6-terra-max	—	71.0	71.0
19	grok-4.7-xhigh	—	69.4	69.4
20	claude-sonnet-5-5-xhigh-effort	—	67.7	67.7
21	deepseek-v4-pro-0813	—	66.1	66.1
22	smaug-flash	—	64.5	64.5
23	gemini-3.1-pro-preview-high	—	62.9	62.9
24	deepseek-v4-flash-vision-exp	—	61.3	61.3
25	smaug-mini	—	59.7	59.7
26	gemini-3.8-flash-high	—	58.1	58.1
27	qwen3.8-flash-next	—	56.5	56.5
28	claude-opus-4-8-max-effort	—	54.8	54.8
29	grok-4.5	—	53.2	53.2
30	claude-opus-4-7-xhigh-effort	—	51.6	51.6
31	glm-5.3	—	50.0	50.0
32	claude-sonnet-5-xhigh-effort	—	48.4	48.4
33	muse-spark-1.1-xhigh	—	46.8	46.8
34	qwen3.8-27b	—	45.2	45.2
35	gemini-3.5-flash-high	—	43.5	43.5
36	gpt-5.2-2025-12-11-high	—	41.9	41.9
37	claude-opus-4-6-thinking-auto-high-effort	—	40.3	40.3
38	deepseek-v4-flash-0731	—	38.7	38.7
39	gemini-3.6-flash-high	—	37.1	37.1
40	qwen3.7-max	—	35.5	35.5
41	gpt-5.2-codex	—	33.9	33.9
42	gpt-5.6-luna-max	—	32.3	32.3
43	claude-sonnet-4-6-thinking-auto-medium-effort	—	30.6	30.6
44	glm-5.2	—	29.0	29.0
45	claude-opus-4-5-20251101-thinking-64k-high-effort	—	27.4	27.4
46	inkling-xhigh	—	25.8	25.8
47	deepseek-v4-pro	—	24.2	24.2
48	gpt-6-luna-max	—	22.6	22.6
49	glm-5.3-flash	—	21.0	21.0
50	kimi-k2.6-thinking	—	19.4	19.4