MB / MODELS

模型排行榜

按日 Token 消耗排序，覆盖 539 个在榜模型

数据日期 2026-09-12 快照数据最近同步 11:04 回到首页

完整榜单 共 539 个模型 · 当前第 8/12 页

#	模型	上下文	日 Tokens	日请求	日变化
340	minimax/minimax-01	—	247.3M	100.1K	17.7%
341	openai/gpt-5-image-mini	—	246.1M	45.9K	1.9%
342	openai/gpt-image-2.5-flare-20260908	—	229.4M	87.5K	—
343	sao10k/l3.1-euryale-70b	—	222M	75.3K	18.4%
344	krea/krea-2-medium-20260720	—	220.5M	51.1K	66.6%
345	qwen/qwen3-vl-235b-a22b-thinking	—	218.3M	43.4K	21.8%
346	openai/gpt-6-astra-20260903 batch	—	211.6M	41.6K	389.0%
347	inclusionai/ling-3.0-flash-vl-20260910	—	206.2M	29.4K	—
348	meta-llama/llama-3.2-1b-instruct	—	191.2M	177.2K	68.5%
349	qwen/qwen3-max-thinking-20260123	—	189.4M	26K	64.1%
350	cohere/command-r-08-2024	—	169.4M	93K	38.8%
351	thedrummer/skyfall-36b-v2	—	164.2M	49.3K	17.5%
352	google/lyria-3-pro-preview-20260330	—	160.7M	45.4K	142.1%
353	openai/gpt-5.6-sol-pro-20260709 batch	—	159.1M	3.7K	162.2%
354	mistralai/mixtral-8x22b-instruct	—	158.4M	230.1K	34.7%
355	upstage/solar-pro-3	—	157M	93K	299.1%
356	qwen/qwen3-next-80b-a3b-thinking-2509	—	153.8M	29.8K	2.8%
357	google/gemini-3.5-flash-20260519 batch	—	147.4M	13.3K	62.2%
358	tencent/hunyuan-a13b-instruct	—	136.9M	30.2K	10.5%
359	qwen/qwen3-vl-30b-a3b-thinking	—	135.4M	32.5K	22.2%
360	mistralai/mistral-saba-2502	—	134.2M	46.6K	10.6%
361	mistralai/voxtral-small-24b-2507	—	129.4M	166.6K	12.9%
362	z-ai/glm-5.3-20260816 batch	—	125.3M	8K	—
363	qwen/qwen-image-3-20260805	—	122.5M	27.8K	10.5%
364	qwen/qwen3-30b-a3b-thinking-2507	—	121.3M	28.1K	77.4%
365	liquid/lfm-2.5-embedding-350m-20260818 free	—	120.8M	96.6K	48.8%
366	anthracite-org/magnum-v4-72b	—	115.6M	37.4K	15.7%
367	undi95/remm-slerp-l2-13b	—	111.8M	60K	2.7%
368	intfloat/e5-large-v2-20251117	—	110.8M	225.6K	8.9%
369	nvidia/nemotron-3-nano-omni-30b-a3b-reasoning-20260428	—	109.2M	5.3K	55.9%
370	openai/gpt-5-image	—	108.9M	26.1K	7.7%
371	cohere/north-mini-code-20260617	—	106.2M	4.9K	471.4%
372	hexgrad/kokoro-82m	—	104.9M	619.5K	31.2%
373	sentence-transformers/paraphrase-minilm-l6-v2-20251117	—	102.4M	24.5K	3203.8%
374	perplexity/sonar-reasoning-pro	—	101.7M	40.7K	19.2%
375	openai/o4-mini-high-2025-04-16	—	100.9M	18.1K	12.3%
376	openai/gpt-4-turbo	—	100.9M	111K	177.6%
377	black-forest-labs/flux.2-max	—	100.7M	10.5K	66.6%
378	perplexity/sonar-deep-research	—	92.3M	5.8K	23.1%
379	deepseek/deepseek-r1-distill-llama-70b	—	91.7M	37.8K	10.6%
380	openai/gpt-6-astra-pro-20260903 batch	—	87.7M	1.5K	3614.9%
381	google/gemma-2-27b-it	—	84M	148.8K	48.6%
382	qwen/qwen3-vl-8b-thinking	—	83.7M	16.2K	19.6%
383	amazon/nova-premier-v1	—	80M	28.1K	3.4%
384	morph/morph-v3-fast	—	78.9M	7.5K	30.3%
385	amazon/nova-pro-v1	—	76.3M	22.4K	21.5%
386	google/gemini-2.5-flash batch	—	72.6M	12.1K	57.9%
387	x-ai/grok-4.3-20260430 batch	—	72.2M	84.8K	13381.7%

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。