# LLM Prices

> Up-to-date LLM API pricing per model: input, output, cache and batch rates, with upcoming price changes tracked. Data sourced from official provider pricing pages.

Site: https://llmprices.org
- [Gemini 3.8 Flash Pricing](https://llmprices.org/m/gemini-3.8-flash-pricing): $0.75/1M input, $3.75/1M output (doubles 2027-01-01) — Google's smartest Flash model, built for long-horizon software engineering, autonomous agents and complex enterprise workflows.
- [Gemini 3.7 Flash Pricing](https://llmprices.org/m/gemini-3.7-flash-pricing): $0.75/1M input, $3.75/1M output (doubles 2027-01-01) — High-efficiency, high-speed Flash model for everyday coding, tool-using agents and stable multi-step workloads.
- [Gemini 3.6 Flash Pricing](https://llmprices.org/m/gemini-3.6-flash-pricing): $0.75/1M input, $3.75/1M output (doubles 2027-01-01) — Previous-generation Flash model balancing speed and multimodal capability for general agent tasks.
- [Claude Opus 5 Pricing](https://llmprices.org/m/claude-opus-5-pricing): $5/1M input, $25/1M output — Anthropic's flagship model for complex, long-horizon tasks.
- [Claude Sonnet 5 Pricing](https://llmprices.org/m/claude-sonnet-5-pricing): $2/1M input, $10/1M output — The best combination of speed and intelligence in the Claude lineup.
- [Claude Opus 4 Pricing](https://llmprices.org/m/claude-opus-4-pricing): $15/1M input, $75/1M output — Previous-generation Claude flagship, still available at premium pricing.
- [Claude Sonnet 4 Pricing](https://llmprices.org/m/claude-sonnet-4-pricing): $3/1M input, $15/1M output — Previous-generation Claude mid-tier model.
- [Claude Opus 5.5 Pricing](https://llmprices.org/m/claude-opus-5.5-pricing): $4/1M input, $20/1M output — Anthropic's latest flagship — a price cut generation: cheaper than Opus 5 while newer.
- [Claude Fable 5.1 Pricing](https://llmprices.org/m/claude-fable-5.1-pricing): $10/1M input, $50/1M output — New Anthropic model family for specialized workloads at premium tier pricing.
- [Claude Mythos 5.1 Pricing](https://llmprices.org/m/claude-mythos-5.1-pricing): $10/1M input, $50/1M output — New Anthropic model family at the premium tier.
- [Claude Haiku 4.5 Pricing](https://llmprices.org/m/claude-haiku-4.5-pricing): $1/1M input, $5/1M output — Anthropic's fastest, cheapest tier — 1/4 the price of Sonnet 5.
- [Claude Opus 4.5 Pricing](https://llmprices.org/m/claude-opus-4.5-pricing): $5/1M input, $25/1M output — Legacy Opus generation — repriced down from $15/$75 to match Opus 5 rates.
- [Gemini 3.1 Pro Pricing](https://llmprices.org/m/gemini-3.1-pro-pricing): $2.0/1M input, $12.0/1M output — Third-generation Pro model for multimodal understanding, agentic capabilities and vibe-coding. Tiered pricing by prompt length.
- [Gemini 3.5 Flash-Lite Pricing](https://llmprices.org/m/gemini-3.5-flash-lite-pricing): $0.3/1M input, $2.5/1M output — Cost-efficient model optimized for high-volume agentic tasks, translation and simple data processing.
- [Gemini 3.1 Flash-Lite Pricing](https://llmprices.org/m/gemini-3.1-flash-lite-pricing): $0.25/1M input, $1.5/1M output — Cost-efficient Flash-Lite for high-volume workloads.
- [Gemini 3.8 Flash TTS Pricing](https://llmprices.org/m/gemini-3.8-flash-tts-pricing): $0.5/1M input, $9.0/1M output (doubles 2027-01-01) — 3.8 Flash text-to-speech audio model engineered for studio-grade voice fidelity and long-form stability.
- [Gemini Omni Flash Pricing](https://llmprices.org/m/gemini-omni-flash-pricing): $1.5/1M input, $9.0/1M output — Next-generation video generation and editing model, generally available on the paid tier.
- [Gemini 3.8 Live Pricing](https://llmprices.org/m/gemini-3.8-live-pricing): $0.75/1M input, $4.5/1M output — Low-latency audio-to-audio model for real-time voice agents and live conversation, with extended-thinking background reasoning.
- [Claude Opus 4.8 Pricing](https://llmprices.org/m/claude-opus-4.8-pricing): $5/1M input, $25/1M output — Late Opus 4 generation — repriced to $5/$25, a third of its original launch price.
- [Claude Opus 4.7 Pricing](https://llmprices.org/m/claude-opus-4.7-pricing): $5/1M input, $25/1M output — Opus 4 generation at the repriced $5/$25 rate.
- [Claude Opus 4.6 Pricing](https://llmprices.org/m/claude-opus-4.6-pricing): $5/1M input, $25/1M output — Opus 4 generation at the repriced $5/$25 rate.
- [Claude Sonnet 4.6 Pricing](https://llmprices.org/m/claude-sonnet-4.6-pricing): $3/1M input, $15/1M output — Late Sonnet 4 generation at $3/$15.
- [Claude Sonnet 4.5 Pricing](https://llmprices.org/m/claude-sonnet-4.5-pricing): $3/1M input, $15/1M output — Sonnet 4 generation at $3/$15.
- [DeepSeek Flash Pricing](https://llmprices.org/m/deepseek-flash-pricing): $0.3/1M input, $1.2/1M output — DeepSeek-V4.1-Flash — flagship value model with 1M context, vision, thinking modes, and off-peak discounts at half price.
- [DeepSeek V4 Pro Pricing](https://llmprices.org/m/deepseek-v4-pro-pricing): $1.32/1M input, $3.96/1M output — DeepSeek-V4-Pro (0813) — the pro tier with 1M context, positioned well below US flagship pricing.
- [GPT-6 Astra Pricing](https://llmprices.org/m/openai-gpt-6-astra-pricing): $10.0/1M input, $50.0/1M output — OpenAI's flagship GPT-6 tier.
- [GPT-6 Sol Pricing](https://llmprices.org/m/openai-gpt-6-sol-pricing): $2.0/1M input, $10.0/1M output — GPT-6 mid tier — the volume workhorse.
- [GPT-6 Luna Pricing](https://llmprices.org/m/openai-gpt-6-luna-pricing): $0.1/1M input, $0.5/1M output — GPT-6 mini tier — 20x cheaper than Astra.
- [GPT-5.6 Sol Pricing](https://llmprices.org/m/openai-gpt-5.6-sol-pricing): $4.0/1M input, $20.0/1M output — GPT-5.6 generation; alias gpt-daybreak-blue-latest points here.
- [GPT-5.6 Cyber Pricing](https://llmprices.org/m/openai-gpt-5.6-cyber-pricing): $12.5/1M input, $75.0/1M output — GPT-5.6 premium tier; alias gpt-daybreak-red-latest points here.
- [GPT-5.3 Codex Pricing](https://llmprices.org/m/openai-gpt-5.3-codex-pricing): $1.75/1M input, $14.0/1M output — Coding-specialized model.
- [GPT Rosalind Research Pricing](https://llmprices.org/m/openai-gpt-rosalind-pricing): $5.0/1M input, $25.0/1M output — Research-oriented model.