Anthropic Claude models through odnoga
The most consistent family on the gateway: every Claude model here reads images, calls tools and carries a cached-input rate. If your workload repeats a long system prompt or a large document, this is where the caching discount pays for itself fastest.
Caching is the whole economic story
Every Claude model on odnoga publishes a cached-input price, and the discount is steep — an order of magnitude off the standard input rate, and steeper still on the newest models. The practical consequence: put the stable part of your prompt first and keep it byte-identical between calls. A long system prompt, a product manual, a schema — these stop being a per-call cost and become close to free after the first request.
Picking a tier
The family is deliberately narrow: a small fast model, a mid model that handles most production work, and a large model for the jobs where a wrong answer is expensive. Most teams over-buy here. Start on the mid tier, measure, and move up only where the evaluation says the larger model actually wins — odnoga lets you change that per route without redeploying, so the decision is reversible.
Vision and tools everywhere
You do not have to check a capability matrix before routing here: image input and tool calling are available across the family, so a fallback from one Claude model to another never silently loses a feature. That is not true across vendors, which is why odnoga checks capability compatibility when it builds a fallback chain.
Vendor cost is the list price. Your price is that plus your plan margin — the same figures that appear on your invoice.
Every model here, cheapest first
| Model | Vendor | Context | In / 1M | Out / 1M | Cached in | Your price in | |
|---|---|---|---|---|---|---|---|
| Claude Haiku 4.5 claude-haiku-4-5 | Anthropic | 200K | $1 | $5 | $0.1 | $1.08 | Details |
| Claude Sonnet 5 claude-sonnet-5 | Anthropic | 1M | $2 | $10 | $0.2 | $2.15 | Details |
| Claude Sonnet 4.5 claude-sonnet-4-5 | Anthropic | 200K | $3 | $15 | $0.3 | $3.23 | Details |
| Claude Sonnet 4.6 claude-sonnet-4-6 | Anthropic | 1M | $3 | $15 | $0.3 | $3.23 | Details |
| Claude Opus 4.5 claude-opus-4-5-20251101 | Anthropic | 200K | $5 | $25 | $0.5 | $5.38 | Details |
| Claude Opus 4.6 claude-opus-4-6 | Anthropic | 1M | $5 | $25 | $0.5 | $5.38 | Details |
| Claude Opus 4.7 claude-opus-4-7 | Anthropic | 1M | $5 | $25 | $0.5 | $5.38 | Details |
| Claude Opus 4.8 claude-opus-4-8 | Anthropic | 1M | $5 | $25 | $0.5 | $5.38 | Details |
| Claude Opus 5 claude-opus-5 | Anthropic | 1M | $5 | $25 | $0.5 | $5.38 | Details |
| Claude Fable 5 claude-fable-5 | Anthropic | 1M | $10 | $50 | $1 | $10.75 | Details |
| Claude Fable 5.1 claude-fable-5-1 | Anthropic | 1M | $10 | $50 | $0.25 | $10.75 | Details |
Your price column is the vendor cost +7%.
Narrow it differently
The widest range on the gateway, and the widest price range with it — from the cheapest token on odnoga to the most expensive.
Models that accept an image alongside the prompt: screenshots, scans, invoices, charts, photographs of paperwork.
Models that take 200,000 tokens or more in a single request — roughly a long book, a full codebase, or a year of support tickets.
Or read how the gateway picks between them: routing and fallback, and what it costs: plans and margins.