LLMs

LLMs ranking

479 published · sponsored rows sit above the ranking

Rank	Model	Developer	Origin	$ / 1M in / out	Context	Open weights	Last changed
Ranking · 479 match
301 —	Phi-3.5-vision-instructMicrosoft model	Microsoft	US	—	—	Yes
302 New	Gemini 3.8 LiveGoogle model	Google	US	—	—	No
303 —	deepseek-moe-16b-chatDeepSeek model	DeepSeek	CN	Free	—	Yes
304 New	Devstral Small 2Mistral AI model	Mistral AI	FR	—	—	Yes
305 —	NVIDIA-Nemotron-Nano-9B-v2NVIDIA model	NVIDIA	US	—	—	Yes
306 New	GPT-5-CodexOpenAI model · 4 versions	OpenAI	US	—	—	No
307 —	GLM 4.5VZhipu AI model · 3 versions	Zhipu AI	CN	$0.60 / $1.80	65K	Yes
308 —	Anthropic: Claude Sonnet 5Anthropic model	Anthropic	US	$2 / $10	1M	No
309 —	Seed-OSS-36B-InstructByteDance Seed model	ByteDance Seed	CN	$9/mo	—	Yes	22 OctHugging Face PRO offers enterprise support and inference endpoints.
310 —	Llama 4 ScoutMeta model	Meta	US	$0.10 / $0.30	1.3M	Yes
311 —	Qwen: Qwen3.7 MaxAlibaba model	Alibaba	CN	$1.48	1M	Yes	21 OctQwen3.7 Max: flagship model for agentic workloads, coding, office tasks. Support
312 —	Phi-4-mini-instructMicrosoft model	Microsoft	US	—	—	Yes
313 New	Gemini Omni FlashGoogle model	Google	US	—	—	No
314 —	DeepSeek R1DeepSeek model	DeepSeek	CN	Free	163K	Yes
315 New	Leanstral 1.5Mistral AI model	Mistral AI	FR	—	—	—
316 —	GPT-5.1-Codex-MaxOpenAI model	OpenAI	US	$1.25 / $10	400K	No
317 —	GLM-4.6V-FlashZhipu AI model	Zhipu AI	CN	—	—	Yes
318 New	Claude Sonnet 5.5Anthropic model	Anthropic	US	$2 / $10	1M	No
319 —	Stable-DiffCoder-8B-InstructByteDance Seed model	ByteDance Seed	CN	Free	—	Yes
320 —	Llama Guard 4 12BMeta model	Meta	US	$0.18 / $0.18	163K	Yes

Latest in LLMs verified changes

Axolotl v0.20.0 release adds GGUF export, Ringmaster context parallelism, and native NVFP4 LoRA.

Axolotl v0.21.0 release adds Reflection's Beam MoE model for QLoRA fine-tuning, native diffusion LM training, and image dataset support.

Gemini 3.1 Pro Preview Custom Tools offers improved tool selection for coding agents and multi-tool workflows. Supports multimodal reasoning

DeepSeek V4.1 Flash introduces DeepInfer Thor TP2 for enhanced text generation capabilities.

DeepSeek V4.1 Flash model released, offering improved IQ and XXS quantization for efficient inference.

LLMs ranking