AI Models

Current OpenAI flagship — advanced reasoning, agentic workflows, and coding (released Apr 23, 2026)

High-reasoning depth variant of GPT-5.5 for the most complex tasks

Previous OpenAI flagship with native computer-use capabilities, still available

Cost-effective GPT-5.4 variant optimised for low latency and production workloads

High-reasoning variant of GPT-5.4 for accuracy-critical enterprise tasks

NEW Mythos-class flagship — Anthropic's most capable model for long-horizon agentic work (released Jun 9, 2026)

Most capable Opus-tier model for complex reasoning and coding (released May 28, 2026)

Powerful Opus model for reasoning-intensive tasks (released Apr 16, 2026)

Balanced Claude model — strong performance and efficiency for everyday tasks (released Feb 2026)

Fastest and most affordable Claude model for high-throughput, low-latency tasks

Premium Gemini 3.5 flagship for advanced reasoning and large-scale multimodal tasks (rolling out June 2026)

High-capability Gemini model for complex reasoning and long-context analysis (2M tokens)

Most cost-efficient and fastest Gemini model — ideal for high-volume, lightweight tasks

Legacy Gemini 2.5 reasoning model — still supported for production workflows

Legacy Gemini 2.5 balanced model — still supported for high-throughput enterprise tasks

Legacy Gemini 2.5 cost-sensitive model — still supported for high-volume operations

xAI current flagship — complex reasoning, multi-step agents, 1M token context ($1.25/1M input tokens)

xAI coding specialist — debugging, web dev, terminal-native workflows at 100+ tokens/sec (released May 2026)

DeepSeek flagship — 1.6T total / 49B active params, 1M token context, thinking + non-thinking modes

Cost-effective DeepSeek model — 284B total / 13B active params, 1M context, agentic workflows