price tracker · updated 2026-07-20
Cheapest Claude Opus 4.8 API
llm · per 1M output tokens
Claude Opus 4.8 is Anthropic's most capable Opus-tier model at $25 per 1M output tokens ($5 input) with a 1M context window at standard pricing; the Batch API runs it for $12.50, 50% cheaper.
Across the 4 endpoints we track for Claude Opus 4.8, posted prices span $12.50 at anthropic (batch) to $25 at amazon bedrock. That is a 2.0x gap for the same model, so provider choice matters more than most teams expect. Rates below are normalized to llm · per 1M output tokens and re-checked weekly.
Claude Opus 4.8
llm · per 1M output tokens
save up to 50%
Cheapest verified endpoint: anthropic (batch) at $12.50 (llm · per 1M output tokens). Prices normalized per unit, ex. egress.
How we track this
We re-pull every known provider that serves Claude Opus 4.8 weekly, normalize to a single unit (llm · per 1M output tokens), and pin the price the moment a provider posts it — we don't average across stale snapshots. Switching providers is usually a one-line base-URL change. Spot a stale price? tell us.
Frequently asked questions
What is the cheapest Claude Opus 4.8 API?
anthropic (batch) currently has the lowest price for Claude Opus 4.8 at $12.50 (llm · per 1M output tokens), based on our 2026-07-20 check across 4 providers.
How much does the Claude Opus 4.8 API cost?
Prices for Claude Opus 4.8 range from $12.50 at anthropic (batch) up to $25 at amazon bedrock, measured as llm · per 1M output tokens. Same model, different bill: picking the right provider saves up to 50%.
Which providers offer Claude Opus 4.8?
We currently track 4 endpoints for Claude Opus 4.8: anthropic (batch), amazon bedrock, openrouter, anthropic. The table above ranks them by effective price, cheapest first.
How often are these prices updated?
We re-check every provider weekly and pin the exact posted price rather than averaging stale snapshots. This table was last refreshed on 2026-07-20.
Cut the bill further
Cheapest AI Inference API in 2026: Ranked by Real Prices
Cost · 6 min read
The Cheapest Way to Run LLMs in 2026 (Without Burning Your Runway)
Cost · 7 min read
Prompt Caching and Batch APIs: Cut Your LLM Bill in Half
Cost · 6 min read
Similar models we track
Cheapest Llama 3.3 70B API
llm · per 1M output tokens · save up to 73%
Cheapest DeepSeek-V3 API
llm · per 1M output tokens · save up to 41%
Cheapest DeepSeek-R1 API
llm · per 1M output tokens · save up to 69%
Cheapest Llama 4 Maverick API
llm · per 1M output tokens · save up to 29%
Cheapest Llama 4 Scout API
llm · per 1M output tokens · save up to 49%
Cheapest Llama 3.1 405B API
llm · per 1M output tokens · save up to 13%
Cheapest Llama 3.1 8B API
llm · per 1M output tokens · save up to 72%
Cheapest Qwen2.5 72B API
llm · per 1M output tokens · save up to 67%
Cheapest Qwen3 235B A22B API
llm · per 1M output tokens · save up to 86%
Cheapest Mistral Nemo API
llm · per 1M output tokens · save up to 76%
Cheapest Mistral Small 3 API
llm · per 1M output tokens · save up to 73%
Cheapest Kimi K2 API
llm · per 1M output tokens · save up to 56%
Browse all 89 model price trackers →
Building on AI? Don't pay full price.
Perkstack also tracks 200+ verified AI tool credits and startup grants, each with a step-by-step claim guide.