ColinBuilds · Legacy Anthropic API model · still available
Claude Opus 4.8
- Input / 1M
- $5.00
- Output / 1M
- $25.00
- Official Anthropic context window
- 1,000,000 tokens
Anthropic Claude API claude-opus-4-8 official API pricing; input $5/1M tokens, output $25/1M tokens, with separate prompt-cache columns. Anthropic lists Opus 4.8 as a legacy model that is still available. Current Opus-tier recommendation is Claude Opus 5. · as of August 25, 2026
Compare models
Same per-1M input/output rates as above. Enter your monthly token volumes to compare bills.
Claude Opus 4.8
- Input: 1M × $5.00 = $5.00
- Output: 1M × $25.00 = $25.00
- Total
- $30.00
GPT-4o
- Input: 1M × $2.50 = $2.50
- Output: 1M × $10.00 = $10.00
- Total
- $12.50
Difference
$17.50
Claude Opus 4.8 costs $17.50 more than GPT-4o.
Popular comparisons
37 total featuring Claude Opus 4.8.
Claude Opus 4.8 is Anthropic’s previous Opus-tier API snapshot. It remains available as a legacy model. Anthropic’s current models overview recommends Claude Opus 5 for complex agentic coding and enterprise work.
What this model is
Anthropic lists the exact Claude API ID and alias as claude-opus-4-8. The models overview now groups Opus 4.8 with other legacy models that are still available, rather than as the current Opus recommendation.
The same overview lists Claude Opus 4.8 with text and image input, text output, multilingual capability, vision support, adaptive thinking, a 1 million token context window, and a 128,000 token maximum output. Anthropic notes that every Claude model ID is a pinned snapshot.
Pricing notes
Anthropic’s official pricing page lists Claude Opus 4.8 at $5 per 1M input tokens and $25 per 1M output tokens, matching Claude Opus 5’s base token rates. The calculator on this page uses those first-party Claude API rates.
The same pricing table lists separate prompt-cache rates for Claude Opus 4.8: 5-minute cache writes at $6.25 per 1M tokens, 1-hour cache writes at $10 per 1M tokens, and cache hits and refreshes at $0.50 per 1M tokens. Anthropic also lists separate pricing paths for Batch API discounts, cloud platforms, data residency, and regional endpoint premiums.
Benchmarks and specs
Anthropic’s models overview lists Claude Opus 4.8 with a 1 million token context window, a 128,000 token maximum output, a reliable knowledge cutoff of January 2026, and a training data cutoff of January 2026.
This page does not show a parameter count or public benchmark score. Those numbers should only be shown when exact public sources are matched to claude-opus-4-8 specifically.
Best fit
Claude Opus 4.8 is best for existing integrations that are pinned to this snapshot, or for historical cost comparisons against the current Opus 5 route.
For new work, Anthropic’s current overview points to Opus 5 as the Opus-tier starting point and Fable 5 as the highest-capability option.