anthropic/claude-haiku-4.5

Anthropic's fastest and most affordable Claude model — ideal for low-latency, high-volume tasks

anthropic/claude-opus-4.6

Anthropic's most intelligent model — state-of-the-art reasoning, extended thinking, 1M-token context, and top-tier coding capabilities

anthropic/claude-sonnet-4.6

Anthropic's flagship production model — 1M context, best balance of intelligence and speed

deepseek/deepseek-r1-distill-qwen-32b

DeepSeek R1 distilled into Qwen 32B — strong reasoning at zero cost on Cloudflare edge

deepseek/deepseek-v3.2

DeepSeek's flagship open-source MoE model — strong reasoning, coding, and math

google/gemini-2.5-flash

Google's fast multimodal model with 1M context, vision, and audio I/O

google/gemini-3-flash-preview

Google's next-generation Gemini 3 Flash — 1M context, enhanced reasoning and multimodal

google/gemini-3.1-flash-lite-preview

Ultra-affordable Gemini 3.1 Flash Lite — very cheap, 1M context, multimodal

google/gemma-3-12b-it

Google's Gemma 3 12B — compact and efficient, free on Cloudflare Workers AI edge

google/gemma-4-26b-a4b-it

Google's Gemma 4 26B — next-generation open model with improved reasoning and efficiency

google/gemma-4-31b-it

Google's Gemma 4 31B — larger open model with stronger reasoning, instruction following, and multilingual performance

meta-llama/llama-3.1-405b-instruct

Flagship open-source model matching frontier performance at 128k context

meta-llama/llama-3.1-70b-instruct

Best open-source model for most production use cases with 128k context

meta-llama/llama-3.1-8b-instruct

Meta's Llama 3.1 8B — lightweight, fast, and free via Cloudflare Workers AI edge

meta-llama/llama-3.3-70b-instruct

Meta's Llama 3.3 70B — high-quality open model, available free on Cloudflare Workers AI

mistralai/mistral-small-3.1-24b-instruct

Mistral Small 3.1 — efficient 24B model with vision, free on Cloudflare Workers AI

openai/gpt-4o-mini

Cost-efficient small model that outperforms GPT-3.5 Turbo on most benchmarks

openai/gpt-4o

Omni model with native vision, audio, and text at GPT-4 intelligence level

qwen/qwen2.5-72b-instruct

Alibaba's Qwen 2.5 72B — strong multilingual reasoning, free on Cloudflare Workers AI

qwen/qwen3.5-9b

Alibaba's Qwen 3.5 9B — compact, efficient, and very affordable reasoning model

stepfun/step-3.5-flash

StepFun Step 3.5 Flash — fast, affordable model with strong multilingual support