Why Providers Differ — Beyond the Marketing

Same task, different shapes

Different providers are not just different models behind the same API. They differ in: tokenizer, role naming, structured-output mechanics, tool-call format, system-prompt semantics, refusal calibration, reasoning configuration, multimodal capabilities, pricing, rate limits, and SDK conventions. The differences add up; a prompt tuned for one rarely lands cleanly on another.

The four families to know in 2026

Anthropic — Claude family. Strong on long context, XML-tagged prompts, extended thinking, tool use.
OpenAI — GPT-5.5 / o-series. Strong on structured outputs (strict JSON schema), reasoning, multimodal.
Google — Gemini 2.5 Pro / Flash. Strong on long context, multimodal, native search grounding.
Open-source — Llama 4, Qwen 3, Mistral, etc. Run anywhere; quality varies; no provider lock-in.

Why this matters

You will move between providers — for cost, for outage hedging, for regional compliance, for capability. Plan for it. The prompt-engineering work is portable; the wiring isn't.

Code

The differences in one table (mental model)·plaintext

                  Claude         OpenAI         Gemini         OSS
roles             system+user    system+dev+u   system+user    varies
XML tag culture   strong         weak           moderate       varies
strict JSON       via tools      json_schema    response_schema partial
reasoning         extended       o-series       thinking        partial
tool_choice       any/specific   required/named auto/any        varies
token counter     server         tiktoken       SDK            varies

Why Providers Differ — Beyond the Marketing

Same task, different shapes

The four families to know in 2026

Why this matters

Code

External links

Exercise

Progress

Comments 0