AI Model Pricing
Input and output rates per 1M tokens for leading OpenAI, Anthropic, and Google Gemini models. Updated April 2026.
OpenAI
GPT-5.4
OpenAI's top-tier model for enterprise — superior reasoning and coding capabilities
GPT-5.4 mini
Most powerful mini model to date — excels at programming, computer use, and subagents
GPT-5.4 nano
Lowest-cost GPT-5.4-tier model — optimized for basic, large-scale operations
GPT-4.1
Prior OpenAI flagship — remains capable in coding, reasoning, and following instructions
GPT-4.1 mini
Streamlined GPT-4.1 — reliable default for most production deployments
GPT-4o
Vision-enabled multimodal model — handles text, images, and audio via single API
GPT-4o mini
Fastest, most affordable GPT-4o variant — suited for high-throughput, low-latency jobs
o3
OpenAI's reasoning specialist — slower yet premium, vastly superior in multi-step logic
o4-mini
Condensed reasoning model — less costly than o3, still outperforms GPT-4o in logic
Google Gemini
Gemini 3.1 Pro
Google's newest flagship — 2M token context, robust in reasoning, coding, and multimodal
Gemini 2.5 Pro
Prior Gemini flagship — 1M token context, extensively used in live systems
Gemini 2.5 Flash
Quick and economical — preferred Gemini for scaled production
Gemini 2.0 Flash
Google's value choice — least expensive Gemini for straightforward, bulk tasks
Anthropic
Claude Opus 4.6
Anthropic's smartest model — ideal for intricate agents, coding, and reasoning challenges
Claude Sonnet 4.6
Prime blend of smarts, value, and pace — Anthropic's standard for most production loads
Claude Haiku 4.5
Swiftest, most economical Claude — perfect for high-volume, low-latency scenarios
Claude Opus 4.1
Legacy high-end Opus — retained for existing integrated workloads
Claude Sonnet 4
Earlier Sonnet release — still common in production environments
Claude Haiku 3
Lowest-price Claude — excellent for basic tasks not requiring newest gen
Contrast all models and project expenses at your scale
Leverage our complimentary tools for side-by-side rates and monthly spend calculations tailored to your scenario.
Compare all models side by side