Skip to content
LLM Toolkit

Comparison hub

Every model, head to head

16 data-driven comparisons across GPT-5, Claude and Gemini — each computed from verified list prices, with monthly cost modeled at three real workloads. Numbers, not vibes.

OpenAI vs Anthropic

GPT-5 vs Claude Opus 5

$1.25 vs $5.00 /1M in · 400K vs 200K ctx

GPT-5 mini vs Claude Sonnet 5

$0.25 vs $3.00 /1M in · 400K vs 200K ctx

GPT-5 nano vs Claude Haiku 4.5

$0.05 vs $1.00 /1M in · 400K vs 200K ctx

GPT-4o vs Claude Sonnet 5

$2.50 vs $3.00 /1M in · 128K vs 200K ctx

OpenAI vs Google

GPT-5 vs Gemini 3.5 Flash

$1.25 vs $1.50 /1M in · 400K vs 1M ctx

Google vs OpenAI

Gemini 3.5 Flash vs GPT-5 mini

$1.50 vs $0.25 /1M in · 1M vs 400K ctx

Google vs Anthropic

Gemini 3.5 Flash vs Claude Sonnet 5

$1.50 vs $3.00 /1M in · 1M vs 200K ctx

Anthropic vs Google

Claude Opus 5 vs Gemini 3.5 Flash

$5.00 vs $1.50 /1M in · 200K vs 1M ctx

OpenAI vs OpenAI

GPT-5 vs GPT-5 mini

$1.25 vs $0.25 /1M in · 400K vs 400K ctx

GPT-5 mini vs GPT-5 nano

$0.25 vs $0.05 /1M in · 400K vs 400K ctx

GPT-5 vs GPT-4o

$1.25 vs $2.50 /1M in · 400K vs 128K ctx

GPT-4o mini vs GPT-5 nano

$0.15 vs $0.05 /1M in · 128K vs 400K ctx

GPT-4o vs GPT-4o mini

$2.50 vs $0.15 /1M in · 128K vs 128K ctx

Anthropic vs Anthropic

Claude Opus 5 vs Claude Sonnet 5

$5.00 vs $3.00 /1M in · 200K vs 200K ctx

Claude Sonnet 5 vs Claude Haiku 4.5

$3.00 vs $1.00 /1M in · 200K vs 200K ctx

Google vs Google

Gemini 3.5 Flash vs Gemini 3.1 Flash-Lite

$1.50 vs $0.25 /1M in · 1M vs 1M ctx

Frequently asked questions

Which LLM is cheapest in 2026?

GPT-5 nano at $0.05 per 1M input tokens and $0.40 per 1M output tokens is the cheapest verified production API. Claude Haiku 4.5 ($1.00 in / $5.00 out) is the cheapest Claude; Gemini 3.5 Flash ($1.50 in / $9.00 out) is Google’s mainstream tier. Every comparison page on this site computes the winner at your real workload.

How do I choose between GPT-5, Claude and Gemini?

Compare on the axes that hit your bill: output price (dominates agent and generation workloads), input price (dominates analysis and RAG), cached-input price (dominates repeat-context apps), and context window (determines what fits in one request). Our head-to-head pages compute all four for each pair, then price three real workloads so you can see dollar differences, not just ratios.

Are these comparison prices up to date?

Yes — every figure comes from our pricing database, where each row carries a verification date checked against official provider pages. All rows on this page were verified as of 2026-08-25. Rows awaiting re-verification are labeled as such rather than silently going stale.

Build your own comparison

Your workload is unique — the API Cost Calculator prices every model at your exact request volumes and cache profile, the Token Counter measures real token usage, and the Pricing Comparison shows all rates on one screen.