基准测试 · TECHAGENT

基准测试是标准化测试，用于评估 AI 模型在推理、知识、数学和编程方面的表现。借助它们可以客观比较模型，并为任务选择合适的模型。

📊 标准. 标准 - 四个最常被引用的公开基准（MMLU、GPQA、HumanEval、SWE-Bench），取自各模型的发布页；我们的评分将其汇总为一个数值。

🤗 HF开放LLM排行榜. HF Open LLM Leaderboard - 对开源模型统一测量的六项任务（IFEval、BBH、MATH、GPQA、MuSR、MMLU-Pro），按平均分排名。

#	模型↕	提供商↕	MMLU↕	GPQA↕	HumanEval↕	SWE-Bench↕	评分↓
1	Claude Opus 4.7	Anthropic	90.1	74.3	92.1	72.5	81.5
2	Moonshot v1 128K	Moonshot AI	83.0	-	77.0	-	79.7
3	GLM-4 Plus	Zhipu AI	83.0	-	76.0	-	79.1
4	ERNIE 4.0 Turbo	Baidu	82.0	-	76.0	-	78.7
5	Doubao Pro 32K	ByteDance	82.0	-	75.0	-	78.1
6	Hunyuan Pro	Tencent	81.0	-	74.0	-	77.1
7	o1	OpenAI	92.3	77.3	92.4	48.9	77.0
8	DeepSeek R1	DeepSeek	90.8	71.5	92.6	49.2	75.1
9	YandexGPT 4 Pro	Yandex	78.0	-	72.0	-	74.7
10	Claude Sonnet 4.6	Anthropic	88.3	65.0	92.0	57.0	74.4
11	MiniMax Text-01	MiniMax	88.0	56.0	84.0	-	73.9
12	o3-mini	OpenAI	86.9	67.4	91.7	49.3	72.9
13	ERNIE 4.5	Baidu	88.0	55.0	82.0	-	72.8
14	GigaChat Max	Sber	74.0	-	68.0	-	70.7
15	DeepSeek V3	DeepSeek	88.5	59.1	89.4	42.0	68.3
16	YandexGPT 3 Pro	Yandex	72.0	-	65.0	-	68.1
17	Qwen Max	Alibaba	89.0	59.0	87.0	37.0	66.5
18	GPT-4o	OpenAI	88.7	53.6	90.2	38.3	65.9
19	HyperCLOVA X	Naver	79.0	-	55.0	-	65.7
20	YandexGPT 4 Lite	Yandex	69.0	-	62.0	-	65.1
21	Llama 3.1 405B	Meta	88.6	50.7	89.0	34.1	63.7
22	GigaChat Pro	Sber	68.0	-	60.0	-	63.6
23	Mistral Large 2	Mistral AI	84.0	47.2	92.1	32.6	62.1
24	Claude Haiku 4.5	Anthropic	82.9	43.0	88.3	33.2	59.9
25	Qwen 2.5 72B	Alibaba/Qwen	86.1	49.0	86.5	23.7	59.5
26	Qwen 2.5 Coder 32B	Alibaba/Qwen	80.0	42.0	92.3	30.0	59.2
27	Gemini 1.5 Pro	Google	85.9	46.2	84.1	26.9	58.8
28	Gemini 2.0 Flash	Google	83.0	45.0	85.0	22.0	56.9
29	Llama 3.1 70B	Meta	83.6	46.7	80.5	21.8	56.3
30	GPT-4o mini	OpenAI	82.0	40.1	87.1	22.8	55.9
31	Gemini 1.5 Flash	Google	78.9	37.0	78.9	16.2	50.7
32	Gemma 2 27B	Google	75.2	38.4	74.0	14.7	48.7
33	Command R+	Cohere	80.4	30.1	74.2	16.8	47.9
34	Mistral NeMo	Mistral AI	68.0	32.0	73.4	15.0	45.3
35	Llama 3.2 3B	Meta	63.4	24.7	58.3	9.5	37.0
36	Claude Opus 4.1	Anthropic	-	-	-	74.5	-

AI模型基准测试