Anthropic: Claude Opus 4.8
Anthropic: Claude Opus 4.8 is anthropic's model. On the un-leaked meo benchmark it ranks #2 of 22 with 70.8% accuracy and an Effective Value (π) of 0.0488 β earning a correct answer for about $0.0584 in 27s each.
as of 2026-06-08
Per-domain accuracy
Efficiency
- Cost per correct
- $0.0584
- Seconds per correct
- 27
- Tokens per correct
- 2,210
- Correct answers
- 156 / 251
- Total benchmark cost
- $9.12
- Avg tokens / item
- 1,373
Robustness under pressure
Susceptibility index -0.031 (Ξaccuracy under threatening/emotional prompts; β0 = robust), sway magnitude 0.031. How we measure this β
- control
- 58%
- neutral
- 46%
- user direct
- 54%
- model direct
- 50%
- user implied
- 58%
- model implied
- 58%
- positive control
- 46%
Aggregated third-party benchmarks
Artificial Analysis (artificialanalysis.ai). Redistribution requires an AA commercial license.
Model details
- Lab
- anthropic
- Released
- 2026-05-27
- Weights
- β
- Context
- 1M
- Max output
- 128K
- Knowledge cutoff
- β
- Price in
- $5.00/M
- Price out
- $25.00/M
- Modality
- text+image+file->text
Sources: openrouter Β· artificial_analysis Β· meo
FAQ
How accurate is Anthropic: Claude Opus 4.8?
Anthropic: Claude Opus 4.8 scores 70.8% overall on the meo benchmark (rank #2 of 22), an un-leaked private holdout graded objective-first.
What is Anthropic: Claude Opus 4.8's Effective Value?
Its Effective Value (π) is 0.0488 at N=10, Ο=1, Ξ΄=1.5 β a fused metric of accuracy, speed, cost, and error-cascade. π rankings shift with task depth.
How much does Anthropic: Claude Opus 4.8 cost per correct answer?
About $0.0584 per correct answer (2,210 tokens, 27 seconds each) at benchmark-time pricing via OpenRouter.
Is Anthropic: Claude Opus 4.8 open weights?
Open-weights status for Anthropic: Claude Opus 4.8 is not recorded.