Claude Opus
Anthropic's flagship Claude Opus model family — Opus 4.8 agentic coding, honesty, Dynamic Workflows; superseded at ceiling by Fable 5
Claude Opus
Entity page for Claude Opus — Anthropic's proprietary flagship model line and a core member of the 2026 "Frontier Trinity."
Current flagship
Claude Opus 4.8 (May 28, 2026) — latest Opus 4.x release before Mythos-class models.
| Attribute | Opus 4.8 |
|---|---|
| Pricing | $5 / $25 per million input/output tokens |
| Context | 1M tokens |
| Release cadence | 41 days after Opus 4.7 |
| Strategic focus | Trustworthy agentic autonomy |
Signature benchmarks (Jun 2026)
| Benchmark | Score | Notes |
|---|---|---|
| SWE-bench Pro | 69.2% | Leads GPT-5.5 by 10.6pp on hard coding |
| SWE-bench Verified | 88.6% | Essentially tied with GPT-5.5 |
| USAMO 2026 | 96.7% | Math leader among Trinity |
| Terminal-Bench | 74.6% | Trails GPT-5.5 |
Key innovation: 4× fewer unreported code flaws vs Opus 4.7; Dynamic Workflows for parallel subagents in Claude Code.
Product ecosystem
| Product | Relationship |
|---|---|
| Claude Code | Agentic coding tool built on Opus; Gartner 2026 Leader |
| Claude Fable 5 | Mythos-class successor for long-horizon coding; falls back to Opus 4.8 on sensitive domains |
| Claude Mythos 5 | Same underlying model, safeguards lifted (gated access) |
Evolution (4.1 → 4.8)
Three phases documented in research:
- Capability foundation (4.1 → 4.5) — coding + reasoning jumps
- Agentic specialization (4.5 → 4.6) — computer use, web search, 1M context
- Reliability & scale (4.6 → 4.8) — honesty, math, Dynamic Workflows
→ Claude Opus Benchmark Evolution 41 To 48 Complete Trend Analysis 2026 05 29
Key articles in this wiki
| Topic | Article |
|---|---|
| Opus 4.8 release | Claude Opus 4 8 Agentic Coding Honesty Dynamic Workflows 2026 05 28 |
| Frontier Trinity | Frontier Trinity Comparison Opus Gpt Gemini Benchmark Showdown 2026 06 01 |
| May showdown | Frontier Showdown May 2026 V4 Gpt55 Opus48 2026 05 29 |
| Fable 5 vs Opus | Claude Fable 5 Mythos 5 Mythos Class Frontier Breakthrough 2026 06 10 |
| Agentic coding tools | Enterprise Ai Coding Agents Showdown Claude Codex Cursor Github 2026 05 27 |
| Fable 5 setup | Howto Claude Fable 5 Agentic Coding Setup |
| Monthly API cost | Ai Coding Agent Monthly Cost Claude Api |
Related
- Organization: Anthropic
- Concept: Frontier Models, Agentic Coding
- Compared to: GPT-5.5, Gemini 3.5 Flash, DeepSeek-V4-Pro, GLM-5.2
Link map
Solid arrows: links from this page. Dashed arrows: pages that link here.
Rendering diagram…