Anthropic builds the Claude line: frontier intelligence with a safety-first split — Fable for everyday work, Mythos for vetted high-security access, Opus and Sonnet for value tiers.
Anthropic is a US AI safety company. Its Claude models lead long-horizon agent work and coding; the 5.1 generation splits one stronger base model into a broad edition (Fable) and a security-first edition (Mythos) for vetted organisations.
| Model | What it is for | Vendor list price |
|---|---|---|
| Claude Fable 5.1flagship · 1 Sep 2026 · 1M ctx | Long-horizon agentic coding and knowledge work; cache reads cut 75%, agent-heavy tasks up to ~45% cheaper than Fable 5 | US$10 / US$50 per 1M (≈ RM41 / RM204) |
| Claude Mythos 5.1security edition · vetted access | Same base as Fable 5.1 with wider safeguards — served only to vetted organisations (trusted-access programme) | US$10 / US$50 per 1M (≈ RM41 / RM204) |
| Claude Opus 5value flagship · 24 Jul 2026 | Near-frontier intelligence at roughly half of Fable-tier pricing | US$5 / US$25 per 1M (≈ RM20 / RM102) |
| Claude Sonnet 5workhorse | Mainstay tier for everyday agent and coding tasks | US$2 / US$10 per 1M (≈ RM8.16 / RM41) |
| Claude Haiku 4.5light · fast | Low-latency classification, extraction and light agent steps | US$1 / US$5 per 1M (≈ RM4.08 / RM20) |
Last updated — Vendor list prices (Sep 2026), approx RM at 1 USD ≈ RM4.08 (spot, 19 Sep 2026). Cache reads on Fable/Mythos 5.1: US$0.25 /1M. Opus 4.5–4.8 remain listed at US$5 / US$25. Figures are vendor public list prices in USD per 1M tokens, standard tier — no batch, cache, promo or peak-hour discount applied. Vendors change list prices without notice: confirm on the vendor’s own pricing page before budgeting.
Discontinued or fully superseded models, kept for historical reference. Active models — even previous-generation ones still on sale — always stay in the lineup above.
| Model | Era & what it was for | Historic list price |
|---|---|---|
| Claude Opus 4.12025 · deprecated | Pre-5-era flagship reasoning model | US$15 / US$75 per 1M (≈ RM61 / RM306) — historic list |
| Claude Sonnet 4 / 4.52025 era | The 2025 workhorse generation before Sonnet 5 | US$3 / US$15 per 1M (≈ RM12.24 / RM61) — historic list |
| Claude Haiku 3.52025 era | Earlier light tier | US$0.80 / US$4 per 1M (≈ RM3.26 / RM16.32) — historic list |
Recent announcements with outlet links (EN first, CN where available).
One stronger base model, two editions. Typical workloads save ~25%, agent-heavy up to ~45% vs Fable 5; Terminal-Bench 4.0 at 55.8%.
EN · VentureBeat ↗EN · Anthropic release notes ↗Anthropic shipped Opus 5 close to Fable-5 intelligence at roughly half the cost, reshaping the value tier.
EN · Anthropic release notes ↗One partner for every digital layer – explore what else we build, host and grow for Malaysian businesses.
