benchrAnthropic
Claude Opus 4.8
Anthropic's deprecation table sends three retired models here: Opus 4 and Opus 4.1, both priced at $15/$75 per million tokens, and Claude 3 Opus, whose retiring rate is not recorded.
The published record
Every number below was read from the provider's own documentation on the date shown. benchr does not restate a figure it has not seen published.
- API identifier
claude-opus-4-8- Context window
- 1MtokensSource
- Maximum output
- 128KtokensSource
- Input
- $5.00per 1M tokensSource
- Output
- $25.00per 1M tokensSource
- Cached input
- $0.5per 1M tokensSource
- Released
- May 28, 2026Source
- License
- Proprietary
Record verified June 12, 2026 — read 86 days ago, and a price can move in a week
What the record says
Standard price unchanged from Opus 4.7. The $10/$50 figure is the optional fast-mode rate (~2.5x output speed), NOT the base price; base is $5/$25. Cache-hit input $0.50 (0.1x base) now explicit on the official pricing page (confirmed June 12, 2026). Official tentative retirement floor: not sooner than May 28, 2027. ARC-AGI-2 is not in the official Opus 4.8 headline summary table, so it is null (do not state one). New Opus 4.7+ tokenizer can use up to ~35% more tokens per the pricing page.
What benchr has documented it doing
Capabilities in the benchr ledger that name this model. Documented means a provider says it works and benchr recorded where. Verified means benchr ran it.
9 capabilities name this model. All 9 are officially supported by the provider; benchr has run 0 of them itself.
- Stop paying twice for the same file
- Operate a desktop application
- Drive a real browser, not a scraper
- Teach the model a procedure once and reuse it
- Read a 600-page PDF including its charts
- Run the whole job overnight at half price
- Let the model write and run real code mid-answer
- Answer from the live web with citations attached
- Work through an entire repository, not a pasted file
What has moved
Entries in the change ledger that affect this model, newest first.
Released
Claude Opus 4.8 launched May 28, 2026. $5/$25 standard; $10/$50 fast mode. SWE-Bench Pro 69.2%. 1M context window confirmed on platform.claude.com. Source: anthropic.com/news
Which one to use
The rest of the family, with the two numbers that usually decide it. The current model is marked.
| Models | Input | Output | Context window |
|---|---|---|---|
| Claude Haiku 4.5 | $1.00 | $5.00 | 200K |
| Claude Sonnet 5 | $2.00 | $10.00 | 1M |
| Claude Sonnet 4.6 | $3.00 | $15.00 | 1M |
| Claude Opus 4.7 | $5.00 | $25.00 | 1M |
| Claude Opus 4.8 This page | $5.00 | $25.00 | 1M |
| Claude Opus 5 | $5.00 | $25.00 | 1M |
| Claude Fable 5 | $10.00 | $50.00 | 1M |
| Claude Fable 5.1 | $10.00 | $50.00 | 1M |
benchr's read
Worth it for
- Highest-stakes coding tasks
- Complex multi-step agents
- Architecture decisions
- Production SWE-bench-level work
Look elsewhere if
- You need cheap volume
- Sonnet handles your workload
- Speed is the priority — use Fast Mode instead
Prices, limits, and identifiers above are provider facts. This section is benchr's judgement about them.
What benchr has written about it
Pieces that name this model, newest first. The ones written about this model come before the ones that mention it in passing.
What is not on this page
Stated rather than filled in.
- No first-token or tokens-per-second figure is recorded. benchr has not measured it and the provider does not publish one.
- benchr has not run any capability on this model itself. Everything in the capability section is documentation, not a test result.