# Claude Opus 4.8

Claude Opus 4.8 is a language model from Anthropic with a 200K-token context window. Provider list price is $5 per million input tokens and $25 per million output; on Allocate you pay $5.35 and $26.75. It is a closed model served over API; the weights are not published.

## Pricing

| | Provider list | On Allocate |
| --- | --- | --- |
| Input, per M tokens | $5 | $5.35 |
| Output, per M tokens | $25 | $26.75 |

Prices checked 2026-07-21.

## Facts

| Field | Value |
| --- | --- |
| Lab | Anthropic |
| Modality | Language |
| Context window | 200K tokens |
| License | Proprietary API |
| Open weights | No |
| Fine-tunable | No |
| Catalog id | anthropic/claude-opus-4.8 |

## What a real workload costs

Take 1,000,000 requests a month at 1,200 input and 350 output tokens each: 1,200M input and 350M output tokens. At list prices that is 1,200 × $5 + 350 × $25 = $14,750 a month. Billed on Allocate it is $15,783.

## Where it fits

Anthropic’s deepest reasoner. At $5 per million input tokens and $25 out it is priced for the hardest problems; route the 1% of traffic where being right is worth it and keep the rest on faster, cheaper models.

- Hardest reasoning problems
- High-stakes analysis
- Escalation tier for agents

## Common questions

### How much does Claude Opus 4.8 cost per million tokens?

Provider list price is $5 per million input tokens and $25 per million output tokens. On Allocate you pay $5.35 in and $26.75 out.

### What context window does Claude Opus 4.8 have?

200,000 tokens (200K). At roughly 0.75 words per token, that is about 150k words of English text per request.

### Can I fine-tune Claude Opus 4.8?

No. Claude Opus 4.8 is a closed model served over API; the weights are not published. If you want a model you can train and own, start from an open-weights base in the catalog and fine-tune that.

### How do I call Claude Opus 4.8 on Allocate?

Send anthropic/claude-opus-4.8 in the model field of the OpenAI-compatible endpoint at api.allocate.network/v1, or point a route name (like prod/support-agent) at it so you can swap the model later without a deploy.

---

[HTML page](https://allocate.network/models/opus-4-8) · [Machine-readable catalog](https://allocate.network/catalog.json)