Claude Opus 5
Last verified July 25, 2026 · Anthropic pricing ↗$5.00/1M input · $25.00/1M output · 1.0M context · Anthropic
Count Tokens
—
Tokens
—
Input Cost
—
Output Cost
Estimate Monthly Cost
Monthly Cost Estimator
Quick:
<$0.0001/mo
Pick a preset above or enter custom usage
Alternatives to Claude Opus 5
Pricing Details
AnthropicAnthropic flagship, launched July 24, 2026. Standard tier $5/$25, identical to Opus 4.8, and exactly half the Fable 5 flagship's $10/$50. Positioned as near-Fable-5 intelligence at half the cost. Fast mode research preview at $10/$50, an exact 2x for 'up to 2.5x higher output tokens per second' - Anthropic states twice that gains are OTPS, NOT time to first token, and publishes no latency or uptime SLA of any kind (contrast OpenAI, which guarantees 99% of Sol Fast requests above 80 tok/s). Fast pricing applies across the FULL context window including requests over 200K input, where OpenAI's Fast excludes long context outright. Fast is gated (account manager or waitlist), first-party API only (not Bedrock, Vertex or Microsoft Foundry), and NOT available with the Batch API; prompt caching does work, with the usual multipliers applying on top of the doubled base, but requests at different speeds do not share cached prefixes. Enabled via speed: 'fast' plus beta header fast-mode-2026-02-01. Priority Tier is NOT supported on Opus 5 (nor Sonnet 5, Mythos 5, Mythos Preview) and Priority Tier capacity commitments are no longer sold, so per-request Fast mode is the only way to buy latency on this model. Cache read $0.50/MTok, cache write (5min/1hr) $6.25/$10, batch $2.50/$12.50 (confirmed on Anthropic's pricing page 2026-08-01). Cache minimum lowered to 512 tokens. Thinking on by default with a low/medium/high/xhigh/max effort ladder; disabling thinking blocked at xhigh/max. Full 1M context flat, no long-context surcharge. SWE-bench Verified ~96.0, SWE-bench Pro ~79.2, ARC-AGI-3 30.2, Frontier-Bench v0.1 43.3 (System Card / Anthropic-reported). AA Intelligence Index ~61 (#1), 53.6 output tok/s on both the Anthropic and Google endpoints (read 2026-08-01); no independent measurement of the Fast tier exists or can exist, since Artificial Analysis's methodology requires the same publicly available configuration any developer receives and explicitly excludes priority queues, and OpenRouter returns null throughput for claude-opus-5-fast. Still behind Claude Mythos 5 on cyber/exploit tasks. API ID: claude-opus-5.
Input / 1M tokens
$5
Output / 1M tokens
$25
Context Window
1.0M
Max Output
128K
Price History
No price change recorded since 2026-07-25, when this site started tracking this model; it launched on 2026-07-24.
Frequently Asked Questions
Common questions about Claude Opus 5 pricing and usage
Read More About Claude Opus 5
DeepSeek kept the price at $0.14 and added ten index points. Eleven models now tie on intelligence, and the bill for proving it runs from $72 to $1,061.
August 2, 2026 · 10 min read
Anthropic shipped Opus 5 at the same $5/$25 it charged for 4.8. The story is what that price now buys: half the Fable 5 bill at close to Fable 5 scores.
July 25, 2026 · 8 min read