LLMs

LLMs ranking

298 published · sponsored rows sit above the ranking

Rank	Model	Developer	Origin	$ / 1M in / out	Context	Open weights	Last changed
Ranking · 298 match
61 —	UncensoredCognitive Computations model	Cognitive Computations	US	$0.20 / $0.90	128K	Yes
62 —	Velvet-14BAlmawave model · 2 versions	Almawave	IT	—	—	Yes
63 New	Xing4.0 29B A4BXingChen model	XingChen		$0 / $0	262K	Yes
64 —	Yi-34B-Chat01.AI model · 3 versions	01.AI	CN	$10/mo	—	Yes
65 —	gpt-oss-120bOpenAI model · 3 versions	OpenAI	US	$0.15 / $0.60	131K	Yes
66 —	llm-jp-3.1-1.8b-instruct4LLM-jp (NII) model	LLM-jp (NII)	JP	—	—	Yes
67 —	oellm-9b-256k-sftOpenEuroLLM model	OpenEuroLLM		—	—	Yes
68 —	plamo-2-1bPreferred Networks model	Preferred Networks	JP	$10/mo	—	Yes
69 —	sabia-7bMaritaca AI model	Maritaca AI	BR	—	—	Yes
70 —	ALLaM-7B-Instruct-previewHUMAIN model	HUMAIN	SA	—	—	Yes
71 —	Apertus-70B-Instruct-2509Swiss AI Initiative model · 4 versions	Swiss AI Initiative	CH	—	—	Yes
72 —	EuroLLM-1.7B-InstructEuroLLM (UTTER) model · 4 versions	EuroLLM (UTTER)		—	—	Yes
73 —	SmolLM2-1.7B-InstructHugging Face model · 3 versions	Hugging Face	US	—	—	Yes
74 —	kanana-1.5-2.1b-instruct-2505Kakao model · 2 versions	Kakao	KR	—	—	Yes
75 —	snowflake-arctic-instructSnowflake model	Snowflake	US	—	—	Yes
76 —	DeepSeek: DeepSeek V3 0324DeepSeek model	DeepSeek	CN	$0.25	163K	Yes	21 OctDeepSeek V3 0324: 685B MoE chat model, 164K context, released Mar 2025. Pricing
77 —	MiniMax: MiniMax M1MiniMax model	MiniMax	CN	$0.55 / $2.20	1M	Yes
78 —	MoonshotAI: Kimi K2 0711Moonshot AI model	Moonshot AI	CN	$0.57	131K	Yes	21 OctKimi K2 0711 by Moonshot AI: 1T params, 32B active. Excels in coding, reasoning,
79 —	NVIDIA: SwitchyardNVIDIA model	NVIDIA	US	—	1M	Yes
80 —	Z.ai: GLM 5.3 FlashXZhipu AI model	Zhipu AI	CN	$0.37 / $1.25	1M	Yes

Latest in LLMs verified changes

Qwen-Plus: 131K context, balanced performance, speed, and cost. Released Feb 1, 2025.

Hermes Agent v0.22.2 fixes a bug where the gateway wouldn't start after a reboot.

Hermes Agent v0.22.4 released with fixes including removing suggestion pills and skipping dependency consent for plugins.

Qwen3.5 Plus 2026-04-20: Updated with tiered pricing above 256K tokens. Accepts text, image, video input.

Qwen3.5 Plus 2026-02-15: Hybrid architecture, improved inference efficiency, state-of-the-art performance in text and multimodal tasks.

LLMs ranking