LLM Reference

LLM Reference is a decision-support directory built for engineers and technology leaders who need to choose the right large language model (LLM) and provider in today's fast-moving AI landscape. It tracks over 1,700 models from more than 130 providers and 235 research labs, with data refreshed weekly to include new releases, verified price changes, and benchmark updates. The core value proposition is simple: stop wasting time hunting through scattered sources and start shipping with confidence. Whether you are building a coding assistant, an agentic workflow, a writing tool, or a research pipeline, LLM Reference gives you a single, trustworthy place to compare models side-by-side, see who offers the cheapest pricing for frontier output, and browse curated editors picks for specific tasks like coding, agents, writing, research, image generation, and video creation. The site is designed for fast triage. You can quickly identify the right model for your job, determine the most cost-effective provider, and get back to building. With a Pulse feed that highlights what changed this week, including new models, price cuts, and benchmark refreshes, LLM Reference keeps you informed without the noise. It is built by the Data Advantage project and updated daily, making it an essential resource for anyone who needs to stay current with the exploding LLM ecosystem. The platform transforms how teams evaluate AI options, turning a chaotic research process into a streamlined, confident decision. It empowers you to move from confusion to clarity, from endless tabs to a single source of truth, so you can focus on what matters most: building great products.

Comprehensive Model Directory

Access a searchable directory of 1,843 language models from 140 providers and 247 research labs. You can search by task, provider, or model name, and filter results to find exactly what you need. The directory includes models for coding, RAG, agents, long context, vision, classification, and JSON or tool use. Each entry includes verified pricing, benchmark scores, and editorial notes, so you can make informed decisions without cross-referencing multiple sources. This feature transforms the overwhelming landscape of LLM options into a manageable, organized catalog that saves hours of research time.

Weekly Pulse Feed

Stay current with the Pulse feed, which highlights everything that changed in the model market each week. This includes 177 new models, 53 price cuts, and 368 benchmark refreshes tracked in a single update. The Pulse feed surfaces the most important changes like new releases such as DiffusionGemma 26B A4B IT and verified provider price reductions. Instead of monitoring dozens of blogs, social media accounts, and research papers, you get a curated summary of what matters. This feature ensures you never miss a critical update that could impact your model selection or budget.

Editors Picks for Specific Tasks

Navigate curated editorial recommendations for six key task categories: coding, agents, writing, research, image generation, and video creation. Each pick comes with a detailed rationale, benchmark scores, and an evaluation of why the model excels for that specific use case. For example, Claude Fable 5 is the top pick for coding with 80.3% SWE-bench Pro and 96% SWE-bench Verified. These picks are researched and updated regularly, so you always have a trusted starting point for any project. This feature eliminates analysis paralysis by giving you expert-backed recommendations that you can implement immediately.

Side-by-Side Model Comparison

Compare any two models directly to see their benchmark scores, pricing, and key capabilities in a single view. The comparison tool includes popular comparisons like Claude Fable 5 versus Claude Opus 4.8, or GPT-5.5 versus Gemini 3.1 Pro Preview. You can also compare pricing across providers to find the cheapest frontier output, with current top pricing at $0.260 per 1M output tokens from Hunyuan HY3 Preview via Tencent Cloud TI Platform. This feature empowers you to make data-driven decisions by putting all relevant information side by side, rather than forcing you to toggle between tabs and manually calculate differences.

Choosing the Best Model for a Coding Assistant

When building a coding assistant, you need a model that excels at code generation, debugging, and understanding complex programming contexts. LLM Reference helps you identify the best coding models by surfacing editors picks like Claude Fable 5, which achieves 80.3% SWE-bench Pro and 96% SWE-bench Verified. You can compare multiple coding models, review their benchmark scores, and check pricing to find the most cost-effective option for your budget. The directory also includes models optimized for specific programming languages and frameworks, so you can fine-tune your selection. Instead of guessing which model will perform best in production, you get verified data that leads to confident, high-quality decisions.

Evaluating Providers for Agentic Workflows

Agentic workflows require models that can handle long tool loops, self-correct without prompting, and maintain context across multiple steps. LLM Reference provides editors picks for agents, such as Claude Sonnet 4.6, which achieves the best generally-available tau-bench score of 87.5. You can compare agent-specific benchmarks across providers, check pricing for production-scale deployment, and review notes on each model's behavior in tool-use scenarios. The platform also tracks which providers offer the best latency and reliability for agentic tasks. This use case transforms a complex evaluation process into a straightforward comparison that saves development teams weeks of trial and error.

Finding Cost-Effective Frontier Models for Research

Research pipelines often require frontier-level performance but must stay within budget constraints. LLM Reference helps you identify the cheapest frontier output, currently $0.260 per 1M output tokens from Hunyuan HY3 Preview. You can browse all providers offering frontier models, compare their pricing structures, and check which benchmarks they excel at for research tasks like summarization, data analysis, and document Q&A. The platform also tracks price cuts weekly, so you can time your deployment to take advantage of the best rates. This use case empowers research teams to maximize their budget without sacrificing model quality, turning cost constraints into strategic advantages.

Selecting Image and Video Generation Models for Creative Projects

Creative teams need models that produce high-quality, brand-consistent visuals and videos. LLM Reference features editors picks for image generation, such as FLUX.2 Dev for photorealistic output with the best text rendering and hands, and for video generation, Veo 3.1 for 30-second clips with native audio up to 4K through Vertex AI. You can compare multiple creative models side by side, review their specific strengths in image editing, music generation, and voice synthesis, and check provider availability. The platform also tracks new creative model releases weekly, ensuring you always have access to the latest capabilities. This use case transforms creative production by providing a trusted, up-to-date resource for selecting the best tools for any visual or audio project.

How often is the model directory updated?

The model directory is updated weekly with new releases, verified price changes, and benchmark updates. The Pulse feed highlights exactly what changed each week, including new models, price cuts, and benchmark refreshes. The platform is built by the Data Advantage project and updated daily, so you always have access to the most current information. This means you can rely on LLM Reference as a single source of truth for the fast-moving LLM landscape.

What types of models are tracked in the directory?

LLM Reference tracks 1,843 language models from 140 providers and 247 research labs. The directory includes models for coding, RAG, agents, long context, vision, classification, JSON or tool use, image generation, video generation, voice synthesis, transcription, and music. Each model entry includes benchmark scores, pricing, and editorial notes. The platform also tracks frontier models, open-weight models, and specialized models for specific tasks, giving you comprehensive coverage of the entire LLM ecosystem.

How are editors picks determined?

Editors picks are curated by the LLM Reference team based on thorough research and analysis of benchmark scores, real-world performance, and community feedback. Each pick includes a detailed rationale explaining why the model excels for its specific task category. For example, coding picks are based on SWE-bench scores, agent picks on tau-bench scores, and writing picks on Chatbot Arena rankings. The picks are reviewed and updated regularly to reflect new model releases and benchmark updates, ensuring you always have current, trustworthy recommendations.

Can I compare pricing across different providers?

Yes, LLM Reference includes a dedicated comparison tool that lets you compare pricing across providers for the same model or for different models. The platform tracks verified price changes weekly and highlights the cheapest frontier output, currently $0.260 per 1M output tokens from Hunyuan HY3 Preview via Tencent Cloud TI Platform. You can also browse all providers to see their pricing structures, including per-token costs for input and output, making it easy to find the most cost-effective option for your specific use case and budget.