模型排行榜

#	模型	上下文	日 Tokens	日请求	日变化
100	google/gemini-2.5-pro	—	56.1B	5.1M	5.8%
101	nvidia/nemotron-3-ultra-550b-a55b-20260604	—	55.8B	2.4M	102.6%
102	V voyageai/rerank-2.5-lite-20260727	—	54.8B	3.4M	619.4%
103	nvidia/nemotron-3.5-lightning-20260807	—	50.3B	2.5M	1.4%
104	N nex-agi/nex-n2.5-mini-20260908 free	—	49.1B	1.2M	—
105	z-ai/glm-4.7-20251222	—	47.5B	3.8M	7.7%
106	openai/gpt-4.1-2025-04-14	—	47.4B	8.6M	26.0%
107	qwen/qwen3-30b-a3b-instruct-2507	—	45.7B	18.8M	7.7%
108	openai/gpt-5.2-20251211	—	45.7B	5M	71.9%
109	deepseek/deepseek-chat-v3.1	—	40.5B	9.6M	3.8%
110	z-ai/glm-5-20260211	—	40.1B	2M	12.4%
111	nvidia/nemotron-3-nano-omni-30b-a3b-reasoning-20260428 free	—	37.4B	1.8M	0.9%
112	qwen/qwen3.5-397b-a17b-20260216	—	36.9B	2.2M	22.6%
113	meta-llama/llama-3.3-70b-instruct	—	33.9B	23.6M	4.9%
114	qwen/qwen3.5-9b-20260310	—	33.5B	8.1M	7.2%
115	qwen/qwen3-vl-32b-instruct	—	31.4B	8.6M	86.4%
116	x-ai/grok-4.20-20260309	—	30.6B	4M	1.7%
117	deepseek/deepseek-chat-v3-0324	—	30.2B	6.1M	9.8%
118	openai/gpt-5.3-codex-20260224	—	29.4B	623.5K	7.5%
119	google/gemini-embedding-2	—	28.8B	18.8M	11.8%
120	qwen/qwen3.8-2.4t-a95b-20260812	—	28.7B	770.5K	31.9%
121	minimax/minimax-m2.5-20260211	—	26.9B	1.6M	1.0%
122	meta/muse-spark-1.2-20260805	—	26.2B	784.6K	62.6%
123	qwen/qwen3-coder-next-2025-02-03	—	26B	1.1M	26.8%
124	qwen/qwen3-32b-04-28	—	24.2B	12.7M	25.7%
125	baai/bge-m3-20251117	—	24.1B	19.5M	20.2%
126	meta-llama/llama-4-scout-17b-16e-instruct	—	22.6B	6.6M	83.3%
127	amazon/nova-micro-v1	—	22.5B	9.9M	3.7%
128	qwen/qwen3-vl-235b-a22b-instruct	—	22.2B	5.3M	16.0%
129	openai/gpt-5.1-20251113	—	21.2B	3M	0.6%
130	mistralai/mistral-small-2603	—	21B	6.2M	8.0%
131	deepseek/deepseek-chat-v3	—	20.7B	8.1M	4.5%
132	perplexity/pplx-embed-v1-0.6B	—	20.2B	8.8M	68.1%
133	qwen/qwen3-coder-480b-a35b-07-25	—	19.5B	836.3K	33.6%
134	qwen/qwen3.6-27b-20260422	—	19.2B	2.2M	51.6%
135	openai/gpt-5.6-terra-pro-20260709	—	19.1B	162.9K	4.4%
136	mistralai/mistral-small-3.2-24b-instruct-2506	—	18.4B	7.2M	12.7%
137	qwen/qwen3.7-max-20260520	—	18.3B	985.2K	0.8%
138	I inception/mercury-2.5-20260908	—	17.1B	1.9M	—
139	qwen/qwen3.6-plus-04-02	—	17B	3.5M	11.1%
140	z-ai/glm-4.7-flash-20260119	—	17B	6.3M	15.3%
141	deepseek/deepseek-v3.2-exp	—	16.8B	2.1M	18.8%
142	meta-llama/llama-4-maverick-17b-128e-instruct	—	16.8B	8.1M	14.3%
143	thinkingmachines/inkling-20260715	—	16.7B	625.7K	66.3%
144	anthropic/claude-4-sonnet-20250522	—	16.1B	2.1M	15.8%
145	anthropic/claude-4.5-opus-20251124	—	14.6B	1.1M	21.8%
146	meta/muse-glimmer-30b-20260810	—	14.3B	2.5M	18.2%
147	mistralai/mistral-small-24b-instruct-2501	—	13.9B	4.2M	8.7%

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。