Workers AI Model Rankings Code Arena

We introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. The model natively processes images and text, and generates text autoregressively.

GLM-5.3 is Z.ai's flagship agentic coding model, pairing a 1M-token context window with reasoning, function calling, and structured outputs to power multi-step, tool-driven development workflows.

Z.ai's flagship agentic coding model

The first natively multimodal model in the GLM-5 series. With 320B total parameters and just 18B active parameters, it outperforms GLM-5.2 across benchmarks and real-world workloads at one-tenth the price, while approaching Claude Opus 4.8 on coding and agentic benchmarks.

Kimi K2.6 is a frontier-scale open-source 1T parameter model with a 262.1k context window, multi-turn tool calling, vision inputs, and structured outputs for agentic workloads.

Kimi K2.7 is a frontier-scale open-source 1T parameter model with a 262.1k context window, multi-turn tool calling, vision inputs, and structured outputs for agentic workloads.

DeepSeek V4 Pro is a high-capability reasoning model from DeepSeek with a one million token context window, built for long-horizon agentic workflows and complex, multi-step problem-solving

DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version, with substantially enhanced agentic capabilities.

GLM-4.7-Flash is a fast and efficient multilingual text generation model with a 131,072 token context window. Optimized for dialogue, instruction-following, and multi-turn tool calling across 100+ languages.

OpenAI’s open-weight models designed for powerful reasoning, agentic tasks, and versatile developer use cases – gpt-oss-120b is for production, general purpose, high reasoning use-cases.

Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Built upon extensive training, Qwen3 delivers groundbreaking advancements in reasoning, instruction-following, agent capabilities, and multilingual support.

Meta's Llama 4 Scout is a 17 billion parameter model with 16 experts that is natively multimodal. These models leverage a mixture-of-experts architecture to offer industry-leading performance in text and image understanding.

SEA-LION stands for Southeast Asian Languages In One Network, which is a collection of Large Language Models (LLMs) which have been pretrained and instruct-tuned for the Southeast Asia (SEA) region.

Clairic is a System One-style model that can extract, compute, estimate, and forecast values from inputs. It is strongly typed and outputs confidence scores and probabilities against its outputs. This model is in early research preview from the Cloudflare Workers AI team.

Clef is a 27B multimodal decision model that turns a state and a schema of typed questions into decisions. It reads the state as text, JSON, images, or video, and returns a probability for every allowed option of every question.

Clef-flash is a fast 9B multimodal decision model that turns a state and a schema of typed questions into decisions. It reads the state as text, JSON, images, or video, and returns a probability for every allowed option of every question.

Clef-omni is a multimodal decision model built on a 30B-parameter mixture-of-experts (3B active) backbone. It turns a state and a schema of typed questions into decisions, reads the state as text, JSON, images, audio, or video, and returns a probability for every allowed option of every question.

PhoneLLM is an open model tuned for voice agents. The model is based on Nemotron Nano 30B base capabilities and excels at typical voice agent tasks with thinking disabled.

DeepSeek-R1-Distill-Qwen-32B is a model distilled from DeepSeek-R1 based on Qwen2.5. It outperforms OpenAI-o1-mini across various benchmarks, achieving new state-of-the-art results for dense models.

This is a Gemma-2B base model that Cloudflare dedicates for inference with LoRA adapters. Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models.

Gemma 4 is Google's most intelligent family of open models, built from Gemini 3 research to maximize intelligence-per-parameter.

This is a Gemma-7B base model that Cloudflare dedicates for inference with LoRA adapters. Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models.

Granite 4.0 instruct models deliver strong performance across benchmarks, achieving industry-leading results in key agentic tasks like instruction following and function calling. These efficiencies make the models well-suited for a wide range of use cases like retrieval-augmented generation (RAG), multi-agent workflows, and edge deployments.

This is a Llama2 base model that Cloudflare dedicated for inference with LoRA adapters. Llama 2 is a collection of pretrained and fine-tuned generative text models ranging in scale from 7 billion to 70 billion parameters. This is the repository for the 7B fine-tuned model, optimized for dialogue use cases and converted for the Hugging Face Transformers format.

Llama 3.1 8B quantized to FP8 precision

The Llama 3.2-Vision instruction-tuned models are optimized for visual recognition, image reasoning, captioning, and answering general questions about an image.

The Llama 3.2 instruction-tuned text only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks.

The Llama 3.2 instruction-tuned text only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks.

Llama 3.3 70B quantized to fp8 precision, optimized to be faster.

Llama Guard 3 is a Llama-3.1-8B pretrained model, fine-tuned for content safety classification. Similar to previous versions, it can be used to classify content in both LLM inputs (prompt classification) and in LLM responses (response classification). It acts as an LLM – it generates text in its output that indicates whether a given prompt or response is safe or unsafe, and if unsafe, it also lists the content categories violated.

The Mistral-7B-Instruct-v0.2 Large Language Model (LLM) is an instruct fine-tuned version of the Mistral-7B-v0.2.

Building upon Mistral Small 3 (2501), Mistral Small 3.1 (2503) adds state-of-the-art vision understanding and enhances long context capabilities up to 128k tokens without compromising text performance. With 24 billion parameters, this model achieves top-tier capabilities in both text and vision tasks.

NVIDIA Nemotron 3 Super is a hybrid MoE model with leading accuracy for multi-agent applications and specialized agentic AI systems.

OpenAI’s open-weight models designed for powerful reasoning, agentic tasks, and versatile developer use cases – gpt-oss-20b is for lower latency, and local or specialized use-cases.

Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen). As of now, Qwen2.5-Coder has covered six mainstream model sizes, 0.5, 1.5, 3, 7, 14, 32 billion parameters, to meet the needs of different developers. Qwen2.5-Coder brings the following improvements upon CodeQwen1.5:

Qwen 3.8 27B is a 27-billion-parameter instruction-tuned language model from Alibaba's Qwen family, designed for vision, efficient general-purpose text generation and agentic workloads.

QwQ is the reasoning model of the Qwen series. Compared with conventional instruction-tuned models, QwQ, which is capable of thinking and reasoning, can achieve significantly enhanced performance in downstream tasks, especially hard problems. QwQ-32B is the medium-sized reasoning model, which is capable of achieving competitive performance against state-of-the-art reasoning models, e.g., DeepSeek-R1, o1-mini.

Apertus 1.5 is an 8B parameter language model designed to advance the state of multilingual, multimodal, fully open, and transparent AI. The models support a wide range of languages, handle contexts of up to 262,144 tokens, and it uses only fully open training data whilst delivering performance comparable to other models of similar size. For access, please fill out this form: https://forms.gle/kgvmkr6ucHN3xNuH6

Inkling 2 Small is a multimodal Mixture-of-Experts model from Thinking Machines Lab with 594B total and 18B active parameters and up to 1M tokens of context. It accepts text and image inputs and generates text, with built-in reasoning and tool calling.

EuroLLM-9B is a 9B parameter model trained on 4 trillion tokens divided across the considered languages and several data sources: Web data, parallel data (en-xx and xx-en), and high-quality datasets. For access, please fill out this form: https://forms.gle/kgvmkr6ucHN3xNuH6