模型排行榜

#	模型	上下文	日 Tokens	日请求	日变化
100	meta/muse-spark-1.2-contributor-20260805	—	60.1B	1.3M	74.1%
101	z-ai/glm-5.1-20260406	—	57.5B	4M	7.1%
102	google/gemini-2.5-pro	—	56.1B	5.1M	2.6%
103	qwen/qwen3.5-35b-a3b-20260224	—	53.7B	6.6M	96.4%
104	nvidia/nemotron-3.5-lightning-20260807	—	48B	2.4M	17.6%
105	openai/gpt-5.2-20251211	—	47.8B	5.6M	62.9%
106	z-ai/glm-4.7-20251222	—	46.2B	3.8M	0.0%
107	openai/gpt-4.1-2025-04-14	—	45.7B	8.6M	27.8%
108	nvidia/nemotron-3-ultra-550b-a55b-20260604	—	45.4B	2.4M	4.1%
109	z-ai/glm-5-20260211	—	42.2B	2.1M	20.8%
110	qwen/qwen3-30b-a3b-instruct-2507	—	42.1B	17.9M	0.2%
111	qwen/qwen3.5-397b-a17b-20260216	—	41.7B	2.7M	52.3%
112	deepseek/deepseek-chat-v3.1	—	39.5B	9M	8.5%
113	meta-llama/llama-4-scout-17b-16e-instruct	—	37.7B	7.4M	209.2%
114	nvidia/nemotron-3-nano-omni-30b-a3b-reasoning-20260428 free	—	36.8B	1.8M	0.8%
115	qwen/qwen3.5-9b-20260310	—	36.1B	8.7M	14.0%
116	minimax/minimax-m2.5-20260211	—	34.6B	1.7M	60.5%
117	qwen/qwen3-vl-32b-instruct	—	34.4B	9.6M	86.2%
118	meta-llama/llama-3.3-70b-instruct	—	32.4B	23.1M	2.7%
119	x-ai/grok-4.20-20260309	—	31.1B	4.2M	4.7%
120	google/gemini-embedding-2	—	30.9B	18.8M	2.5%
121	I inception/mercury-2.5-20260908	—	30.7B	4.2M	—
122	openai/gpt-5.3-codex-20260224	—	30.3B	620.3K	11.8%
123	deepseek/deepseek-chat-v3-0324	—	29.2B	6.2M	2.5%
124	thinkingmachines/inkling-20260715	—	29B	860.9K	22.7%
125	qwen/qwen3.8-2.4t-a95b-20260812	—	25.4B	711.3K	38.0%
126	qwen/qwen3-coder-next-2025-02-03	—	24.5B	987.8K	11.7%
127	baai/bge-m3-20251117	—	23.6B	20.5M	18.1%
128	meta/muse-spark-1.2-20260805	—	23.3B	1M	62.7%
129	perplexity/pplx-embed-v1-0.6B	—	22.4B	8.9M	42.2%
130	qwen/qwen3-32b-04-28	—	22.4B	11.4M	33.1%
131	qwen/qwen3-vl-235b-a22b-instruct	—	22.3B	5.4M	15.4%
132	openai/gpt-5.1-20251113	—	22B	3.2M	2.2%
133	amazon/nova-micro-v1	—	22B	10.3M	11.6%
134	mistralai/mistral-small-2603	—	21.9B	5.8M	15.5%
135	deepseek/deepseek-chat-v3	—	20.6B	7.7M	6.7%
136	mistralai/mistral-small-3.2-24b-instruct-2506	—	20.3B	7.8M	32.6%
137	qwen/qwen3.6-27b-20260422	—	19.2B	2.2M	39.1%
138	qwen/qwen3.7-max-20260520	—	18.7B	1.1M	8.1%
139	deepseek/deepseek-v3.2-exp	—	17.5B	2.4M	2.7%
140	z-ai/glm-4.7-flash-20260119	—	17.4B	6.5M	2.0%
141	openai/gpt-5.6-terra-pro-20260709	—	16.9B	145.9K	23.1%
142	meta-llama/llama-4-maverick-17b-128e-instruct	—	16.5B	7.9M	16.1%
143	anthropic/claude-4-sonnet-20250522	—	16.4B	2.2M	11.4%
144	qwen/qwen3-coder-480b-a35b-07-25	—	16.4B	781.2K	2.8%
145	meta/muse-glimmer-30b-20260810	—	16.3B	3.1M	44.8%
146	qwen/qwen3.6-plus-04-02	—	16.2B	3.3M	12.6%
147	qwen/qwen3.6-flash	—	14B	2.2M	22.5%

口径：日消耗 = prompt + completion + reasoning tokens；变化为相对前一日的比例。页面尚未公开来源链接，数据仅供参考。