Parameters
-
Context Length
1M
Modality
Multimodal
Architecture
Dense
License
Proprietary
Release Date
28 May 2026
Knowledge Cutoff
-
Rank
#26
| Benchmark | Score | Rank |
|---|---|---|
Mathematics max | 0.94 | 6 |
Agent Arena high | 0.10 | 7 |
Reasoning max | 0.89 | 7 |
General Text high | 1482 | ⭐ 9 |
General Text Standard | 1474 | ⭐ 18 |
Coding max | 0.82 | 9 |
Web Development high | 1564 | ⭐ 12 |
General max | 0.76 | 14 |
LiveBench Average max | 0.76 | 18 |
Agentic Coding max | 0.51 | 19 |
Data Analysis max | 0.66 | 38 |
Overall Rank
#26
Coding Rank
-
Anthropic's flagship model released May 28, 2026. Built on the Opus 4 architecture with significant improvements in coding agent performance and judgment quality. Opus 4.8 is approximately four times less likely to allow flaws in code it has written to pass unremarked compared to its predecessor. It introduces Dynamic Workflows in Claude Code, enabling hundreds of parallel subagents in a single session for codebase-scale migrations. Features a 1M token context window, adaptive thinking support, and is priced at $5/M input tokens and $25/M output tokens.
Detailed architecture specifications are limited for proprietary models.
The Claude Opus 4.8 generation represents Anthropic's most capable Opus-class model, built for complex reasoning, long-horizon agentic coding, and high-autonomy professional work. It introduces a major leap in judgment quality and multi-agent coordination, with the ability to run hundreds of parallel subagents simultaneously through the Dynamic Workflows feature in Claude Code.
APX AI
Online