AI Models

Anthropic's coding- and agent-focused flagship in the Claude 4 family.

OpenAI's next-generation flagship model succeeding GPT-4o.

xAI's Grok 4 reasoning flagship.

Claude Opus 4 is Anthropic's top-tier model from the Claude 4 generation, released in May 2025 alongside Sonnet 4. It is built for long, autonomous coding and agent work, and can think at length while using tools.

OpenAI o3 is the successor to o1 in OpenAI's reasoning line. Released in April 2025, it can call tools such as web search and Python while it reasons, and it reasons over images as well as text.

API-only OpenAI model focused on coding and instruction following, with a context window of about 1M tokens.

Meta's Llama 4 MoE family with very long context and multimodal input.

Reasoning-focused member of the Gemini 2.5 family.

Anthropic's first hybrid reasoning model, offering both standard responses and an extended thinking mode. Launched alongside Claude Code.

Open-weight reasoning model that popularized reinforcement-learned chain-of-thought at scale.

671B MoE open-weight base model.

Google's fast, low-cost Gemini 2.0 model with native tool use and a 1M-token context window; generally available from February 2025.

Open-weight 70B text model from Meta that Meta says approaches Llama 3.1 405B quality at a fraction of the serving cost.

OpenAI o1 is a reasoning model that thinks through a problem internally before replying, trading latency and cost for much higher accuracy on maths, science and difficult programming tasks. It reached general availability in December 2024 and set the template for later reasoning models.

Mistral's 123B flagship model with strong multilingual and coding ability; weights released under a research licence.

OpenAI's small, inexpensive multimodal model that replaced GPT-3.5 Turbo as the default low-cost option.

Mid-tier 3.5 model that outperformed Claude 3 Opus on many tasks.

Omni-model unifying text, vision and audio in a single network with realtime capabilities.

8B and 70B open-weight release, later extended to Llama 3.1 and 3.2 with vision.

Top-tier model in the Claude 3 lineup with strong reasoning.

Introduced the 1M–2M token context window on a mixture-of-experts architecture.

Sparse mixture-of-experts open-weight model from Mistral.

First Gemini generation with Ultra, Pro and Nano sizes.

First Grok model from xAI, later open-sourced.

Compact 7B open-weight model that punched above its parameter count.

First widely-licensed open-weight Llama release.

100K-token context refresh of the Claude line.

Multimodal successor to GPT-3.5 with substantial gains on reasoning benchmarks.

First public Claude model from Anthropic.

Instruction-tuned refresh of GPT-3 that powered the original ChatGPT launch.

175B parameter model that established the modern LLM API paradigm.

Second-generation GPT model, initially released with a staged rollout over safety concerns.