Opus 4.8 Same Price, Independent #1 — Reliability and Unit Economics Over Raw Peak
Claude Opus 4.8 ships at Opus 4.7’s $5/$25 with dynamic workflows, cheaper fast mode, and effort controls. Artificial Analysis put it #1 overall at launch (61.4) — Anthropic reclaiming the public intelligence crown from GPT-5.5 until Fable 5. Mythos stays Glasswing-gated.
Same-price upgrade plus an independent #1 is why May 28 mattered beyond feature bullets. Anthropic competed on reliability and unit economics, not only raw capability — and reclaimed the public intelligence crown from OpenAI’s April GPT-5.5 peak until Fable 5 arrived in June.
Anthropic (May 28) launched Claude Opus 4.8 (claude-opus-4-8) at the same standard price as Opus 4.7 — $5 / $25 per million tokens. Fast mode: 2.5× speed at $10 / $50 — three times cheaper than prior Opus fast pricing. Companion features: dynamic workflows in Claude Code (hundreds of parallel subagents), effort controls on claude.ai and Cowork, mid-conversation system messages on the Messages API. Mythos-class models remain limited under Project Glasswing until stronger cyber safeguards allow broader release.
Model and pricing
| Surface | Detail |
|---|---|
| Standard | $5 / $25 per MTok (unchanged) |
| Fast mode | 2.5× speed at $10 / $50 (3× cheaper than prior Opus fast) |
| Availability | Claude products + API; Amazon Bedrock, Google Cloud, Microsoft Foundry |
| Context | 1M-token window by default on those cloud platforms (API release notes) |
Anthropic called Opus 4.8 a “modest but tangible” improvement. Internal evals: ~4× less likely than Opus 4.7 to let code flaws pass unremarked; lower deception rates; new highs on prosocial / user-autonomy measures.
Claims vs checks
| Source | Snapshot | Standing |
|---|---|---|
| Artificial Analysis Intelligence Index | AA post May 28, 2026 | Opus 4.8 = 61.4 — #1 overall at launch, +4.1 vs Opus 4.7, +1.2 vs prior leader GPT-5.5 (xhigh) |
| GDPval-AA | Same | ~1890 Elo — Anthropic retakes #1 agentic knowledge-work board |
| Appwrite Arena | June 2026 board update | #1 without-skills at 97.4% — first above 97% and first to clear Opus 4.7 on that board |
| Arena Text Overall | Mid-July boards | Opus 4.8 thinking variants remain top-5 preference after Fable 5 took absolute #1 |
Lab honesty/alignment rates are vendor. AA and Appwrite are independent snapshots — treat Elo and Index scores as time-stamped boards, not permanent crowns.
Limits
- “Four times less likely” flaw-pass claim is Anthropic internal, not third-party audited.
- Dynamic workflows are research preview (Enterprise/Team/Max).
- Mythos remains gated — general-access vs dual-use restricted split continues.
Sources
- Anthropic: “Introducing Claude Opus 4.8” (May 28, 2026).
- Claude Opus 4.8 System Card (May 28, 2026).
- Artificial Analysis: “Claude Opus 4.8 - The new #1 AI model” (May 28, 2026).
- Appwrite Arena June 2026 leaderboard update.
- Anthropic API release notes (May 28, 2026) for model ID, context, mid-conversation system messages, effort defaults.