Open Source Model

Claude Haiku 4.5

✓

Fast and affordable coding model

Released October 15, 2025, Claude Haiku 4.5 is Anthropic's latest small model that delivers similar coding performance to Sonnet 4 at a third of the cost and more than twice the speed. This generation of Haiku is the first to include features previously exclusive to higher-tier models, including extended thinking capabilities, computer use, and enhanced context awareness.

Claude Sonnet 4.5

✓

The best coding model in the world

Released September 29, 2025, Claude Sonnet 4.5 is the best coding model in the world, achieving 77.2% on SWE-Bench Verified benchmark, outperforming competitors like GPT-5. It's the strongest model for building complex agents and the best model at using computers. Features extended thinking mode for complex problem-solving and is Anthropic's most aligned frontier model for safety. Shows substantial gains in reasoning and math.

DeepSeek Coder

✓

Open-source coding LLM with 236B parameters and ultra-affordable API pricing

DeepSeek Coder is a powerful open-source coding LLM with 236 billion parameters and a 128K context window. Released in June 2024 (V2), it supports 338+ programming languages with an MoE (Mixture of Experts) architecture. DeepSeek offers exceptional performance matching GPT-4 Turbo at extremely affordable prices ($0.14-0.28 per million tokens). Perfect for budget-conscious developers and teams who need powerful AI coding capabilities without breaking the bank.

Gemini 3 Pro

✓

State-of-the-art reasoning and multimodal model

State-of-the-art reasoning and multimodal understanding model with powerful agentic and coding capabilities.

GPT-4o

✓

Faster and more affordable GPT-4

A multi-modal model that's 2x faster, 50% cheaper and has 5x higher rate limits compared to the latest GPT-4 Turbo release.

GPT-5

✓

Next-generation multimodal AI model

A multimodal large language model developed by OpenAI, it is a system that contains a fast, high-throughput model, a deeper reasoning model, and a real-time router that decides which model to use based on conversation type, complexity, tool needs, and explicit user intent.

StarCoder 2

✓

State-of-the-art open-access code LLM by BigCode

StarCoder 2 is a family of open-access Large Language Models for code, developed by BigCode (Hugging Face & ServiceNow). Released in February 2024, it comes in 3B, 7B, and 15B parameter sizes. Trained on The Stack v2 (4 trillion tokens), it offers state-of-the-art performance, transparency, and a 16K context window, making it a powerful tool for code completion and generation.