Anthropic has shipped four Claude models in under seven weeks — Mythos 5 and Fable 5 on June 9, Sonnet 5 on June 30, and Opus 5 on July 24, 2026. For Indian developers, startups, and enterprise teams already budgeting AI spend in dollars, that pace makes one question urgent: which model should you actually be running, and for what?
Short answer: Sonnet 5 for everyday, high-volume work; Opus 5 for daily-driver agentic coding and business workflows where quality matters more than raw token cost; Fable 5 only when a task genuinely needs the highest reasoning ceiling Anthropic offers. Here’s how the three compare, model by model.
Quick comparison
| Model | Price (in/out per 1M tokens) | Context window | Speed / effort control | Best for |
| Sonnet 5 | $2 / $10 (intro, till Aug 31); $3 / $15 after | 1M tokens, 128K output | Fast; effort levels incl. xhigh | High-volume production work, everyday coding, cost-sensitive agents |
| Opus 5 | $5 / $25 (same as Opus 4.8); Fast mode at 2x | 1M tokens, 128K output | Configurable low/med/high/xhigh/max effort; Fast mode ~2.5x speed | Daily-driver agentic coding, long-running tasks, business workflows |
| Fable 5 | $10 / $50 | 1M tokens, 128K output | Highest ceiling, slower, more tokens burned | Hardest reasoning, multi-day autonomous agents, frontier coding |
At current exchange rates (~₹96.5/USD as of late July 2026), that price gap is meaningful for India-based teams: a million tokens on Sonnet 5 costs roughly ₹190–965 depending on input/output mix, versus ₹480–2,415 on Opus 5 and ₹965–4,825 on Fable 5. For teams billing in dollars via the API, the spread compounds fast at production volume.
Claude Opus 5: the new daily driver
Opus 5 launched on July 24 as, in Anthropic’s own framing, a model that comes close to Fable 5’s frontier intelligence at half the price. It’s priced identically to its predecessor Opus 4.8 — $5 per million input tokens, $25 per million output — but delivers what Anthropic calls a step change in coding and knowledge-work benchmarks.
Where Opus 5 stands out:
- It’s now the default model on Claude Max and the strongest model available on Claude Pro.
- A configurable effort setting (low through max) lets teams trade intelligence for token cost on the fly.
- Fast mode runs about 2.5x the default speed, at twice Opus 5’s base price.
- Anthropic’s own testing found it to be the most aligned Claude model released so far, with the lowest rates of deceptive or reckless behaviour among its recent models.
Anthropic is explicit that Opus 5 does not push the frontier on “risky, dual-use” capabilities like offensive cybersecurity or autonomous biology research — that ceiling stays with Mythos 5. For everyday business and engineering use, though, Opus 5 is positioned as the model most teams should reach for first.
Claude Fable 5: the ceiling, not the daily default
Fable 5 launched June 9 alongside Mythos 5 as Anthropic’s top publicly available tier, priced at $10 per million input tokens and $50 per million output. It briefly went offline on June 12, 2026, after a U.S. Department of Commerce export-control directive; access was restored on July 1 once those controls were lifted. Anthropic’s account of the episode is at anthropic.com/news/fable-mythos-access.
Fable 5 still carries the highest reasoning ceiling in the public Claude lineup and remains Anthropic’s recommendation for the hardest, longest-horizon agentic work — tasks a model might run autonomously for days. But it also draws criticism for its token burn rate: it tends to use more tokens per task than Sonnet 5 or Opus 5, which can blow through budgets faster than the sticker price suggests. For most day-to-day work, Opus 5 now gets close enough at half the cost that Fable 5 becomes a tool for the genuinely hardest 10–20% of tasks rather than a default.
Claude Sonnet 5: the volume workhorse
Sonnet 5 launched June 30 as an upgrade to Sonnet 4.6, aimed at production agents, high-volume pipelines, and everyday professional work — document drafting, spreadsheet analysis, presentations, and agentic coding. It carries the same 1M-token context window and 128K max output as Opus 5 and Fable 5, which is new for the Sonnet tier.
Its introductory pricing of $2 per million input tokens and $10 per million output tokens holds through August 31, 2026, after which it moves to standard pricing of $3/$15. It’s also the new default model for Free and Pro plan users in Claude Code and the Claude apps, while Max, Team Premium, and Enterprise pay-as-you-go accounts default to Opus 4.8 (switchable to Opus 5).
For Indian teams running high-volume content pipelines, customer support agents, or large-scale document processing — where the marginal cost per call matters more than squeezing out the last few points of benchmark performance — Sonnet 5 is the pragmatic default.
Which model should you actually use?
- Cost-sensitive, high-volume, or real-time work (support bots, content pipelines, routine coding): Sonnet 5.
- Daily agentic coding, business workflows, financial or legal analysis, and general-purpose “I need this done well” tasks: Opus 5.
- Multi-day autonomous agents, the hardest reasoning or research problems, or frontier coding work where Opus 5 falls short: Fable 5, deployed selectively given its cost.
A practical pattern several early-access partners described: run Opus 5 as the default and route only the tasks it can’t handle — or where verification matters most — up to Fable 5. Anthropic’s own automatic-fallback feature, released alongside Opus 5, is built for exactly this kind of routing.
FAQs
Is Claude Opus 5 better than Fable 5?
Not across the board. On several coding benchmarks Opus 5 comes within a fraction of a percentage point of Fable 5’s peak scores at roughly half the cost per task, and it leads on some knowledge-work evaluations. Fable 5 still holds the edge on the hardest, longest-horizon autonomous work.
Is Sonnet 5 good enough to skip Opus 5 entirely?
For high-volume, cost-sensitive workloads, yes for many teams — Sonnet 5 offers near-Opus quality on coding and agentic benchmarks at a fraction of the price. For work where output quality directly affects revenue or risk, Opus 5’s added reliability tends to justify the higher cost.
Why was Fable 5 briefly unavailable?
Anthropic suspended access to Fable 5 and Mythos 5 on June 12, 2026, to comply with U.S. Department of Commerce export controls. The Department lifted those controls on June 30, and Anthropic restored access on July 1, 2026.
What does Opus 5’s effort toggle actually do?
It lets developers choose between low, medium, high, xhigh, and max reasoning effort per request — trading intelligence for speed and token cost, so the same model can be tuned for a quick reply or a hard multi-step problem.
Sources: Anthropic, “Introducing Claude Opus 5” (anthropic.com/news/claude-opus-5); Anthropic Claude Sonnet page (anthropic.com/claude/sonnet); Anthropic Claude Platform docs.
Sethuram Kishore is the founder and editor of Worthview, an online publication established in 2008. With over 18 years of experience in SEO, digital marketing, and online publishing, he writes about AI, technology, business, and digital trends. He is also the founder of MoneyHulk, a personal finance and business publication.
