Models

Top AI models to track.

Frontier, open-weight, coding, multimodal, and specialist models with practical notes and links to deeper information.

01 / OpenAI

GPT-5.5

OpenAI frontier model used for complex coding, research, agentic workflows, and real-world professional work.

Deep reasoningCodingResearch

02 / OpenAI

GPT-5.4

Strong general-purpose OpenAI model for everyday coding, analysis, writing, and application workflows.

CodingAnalysisTool use

03 / OpenAI

GPT-5.3 Codex

Coding-optimized OpenAI model focused on repository work, implementation, refactors, and debugging.

Code generationRepo navigationDebugging

04 / OpenAI

GPT-4o

A multimodal general-purpose model family designed for fast text, vision, and audio-native assistant experiences.

Multimodal UXGeneral reasoningTool use

05 / Anthropic

Claude 3.5 Sonnet

A strong general assistant and coding model known for writing quality, analysis, and long-context workflows.

WritingCodingLong-context analysis

06 / Anthropic

Claude 3 Opus

Anthropic model known for high-quality reasoning, writing, and complex analysis across long documents.

ReasoningWritingLong documents

07 / Google DeepMind

Gemini 3 Pro

Google's preview flagship Gemini model for advanced reasoning, multimodal understanding, coding, and agentic workflows.

Advanced reasoningMultimodal understandingCoding

08 / Google DeepMind

Gemini 3 Flash

Google's balanced Gemini 3 model built for speed, scale, frontier intelligence, and multimodal application workflows.

SpeedScaleMultimodal tasks

09 / Google DeepMind

Gemini 2.5 Pro

Google's stable thinking model for complex coding, reasoning, STEM, long documents, and multimodal analysis.

ThinkingCodingLong context

10 / Google DeepMind

Gemini 2.5 Flash

Google's price-performance Gemini model for low-latency, high-volume tasks that still benefit from thinking.

Low latencyCost efficiencyThinking

11 / Google DeepMind

Gemini 1.5 Pro

Legacy long-context Gemini model for documents, code, video, and structured inputs. Deprecated in the Gemini API after September 2025.

Long contextMultimodal analysisGoogle ecosystem

12 / Google DeepMind

Gemini 1.5 Flash

Legacy fast Gemini model for lower-latency and cost-sensitive multimodal workloads.

SpeedCost efficiencyMultimodal tasks

13 / Meta

Llama 3.1

An open-weight model family used widely for self-hosted assistants, fine-tuning, and application-specific deployments.

Open weightsFine-tuningSelf-hosting

14 / Meta

Llama 3.1 405B

Large open-weight Llama model used for high-end self-hosted and fine-tuned applications.

Open weightsFine-tuningSelf-hosting

15 / Mistral AI

Mistral Large

A frontier-style commercial model from Mistral focused on multilingual reasoning, coding, and enterprise deployment.

MultilingualEnterprise deploymentCoding

16 / Mistral AI

Mixtral 8x22B

Sparse mixture-of-experts open model used for efficient high-capability text generation and reasoning.

Open model ecosystemEfficiencyMultilingual

17 / DeepSeek

DeepSeek Coder

A coding-focused model family popular for code generation, repository tasks, and open model experimentation.

Code generationOpen model ecosystemDeveloper workflows

18 / DeepSeek

DeepSeek-V2

Efficient mixture-of-experts model family used for general text, code, and cost-sensitive deployment.

EfficiencyCodingOpen ecosystem

19 / Alibaba Cloud

Qwen 2.5

Open model family with strong multilingual, coding, and general assistant capabilities across sizes.

MultilingualCodingOpen weights

20 / Cohere

Command R+

Enterprise-focused model optimized for retrieval-augmented generation, tool use, and business workflows.

RAGEnterprise searchTool use

21 / xAI

Grok

xAI assistant/model family integrated with X and positioned around real-time information access.

Realtime contextConsumer assistantSocial signal

22 / Stability AI

Stable Diffusion 3

Image generation model family focused on prompt adherence, typography, and open creative workflows.

Image generationCreative workflowsOpen ecosystem

23 / Runway

Runway Gen-3

Video generation model family for cinematic generation, editing, and creative production workflows.

Video generationCreative productionEditing workflows