Model landscape — October 2026

There is no universal “number one” any more: each family has its own usage profile. The right reflex is to match the model to the task.

Family	Provider	Strengths	Recommended use
Claude Opus 5.5 / Sonnet 5.5 / Haiku 5.5	Anthropic	Rigorous long-instruction following, code, reliable multi-step agents.	Professional writing, software development, agents.
GPT-6.x	OpenAI	General versatility, broad tool ecosystem (ChatGPT, Codex), good intent inference.	General assistant, automation.
Gemini 3.x — Gemini 4 expected	Google DeepMind	Multimodality, very large contexts, Workspace integration.	Large documents, multimodal research.
Mistral Large / Medium / Small	Mistral AI	European hosting, value for money, data sovereignty.	GDPR or data-residency constraints.
Open models Llama, DeepSeek, Qwen, Kimi…	Various	Self-hosting, fine customisation, controlled cost at high volume.	R&D, sensitive cases needing local hosting.

Checked on 27 October 2026. Versions change almost monthly: consult each provider’s official page before any deployment.

Legacy: from Davinci to chat APIs

text-davinci-003 and other GPT-3-era completion models have been retired since 2024; GPT-NEO, T5, MT-NLG or Wu Dao 2.0 are now history or research material. For any new project use a chat-style API (system / user roles). Example to adapt with the current model ID:

from openai import OpenAI
client = OpenAI()
r = client.chat.completions.create(
    model="<model-id>",
    messages=[
        {"role": "system", "content": "You are an assistant specialised in prompt engineering."},
        {"role": "user", "content": "Which parameters should I set for a reliable prompt?"}],
)
print(r.choices[0].message.content)

Depending on the model, some settings (temperature, top_p) are fixed or replaced by a reasoning-effort setting.

Anthropic — Claude modelsVersions, contexts, pricing.
OpenAI — ModelsCatalogue and deprecations.
Google — Gemini modelsIDs and limits.
Mistral — ModelsMistral catalogue.
Hugging Face — DocumentationOpen models, GPT-NEO, T5, etc.