18 September 2026 · AI & LLMs · 6 min read
When Claude Sonnet 5 launched, Anthropic billed $2 per million input tokens and $10 per million output as introductory pricing — through August 31, 2026, with a step up to $3/$15 scheduled for September 1. I went looking for that increase to update our own pricing page, because it should have landed over two weeks ago. It didn't. Anthropic's official pricing docs now carry a note saying the hike "will not occur" and the $2/$10 launch price is permanent. So I pulled the full Claude 5 family rate card — Fable 5, Opus 5, Sonnet 5, Haiku 4.5 — straight from the source and ran real workloads through it to see what that cancellation is actually worth, and how the four tiers compare once you're not just reading a table.
Take a mid-size product: a support copilot doing 2,000 requests a day, each carrying roughly 8,000 tokens of retrieved context and conversation history in, and returning a 500-token answer. That's 480 million input tokens and 30 million output tokens a month — not a huge company, just a real feature with real traffic.
| Sonnet 5 rate | Input (480M) | Output (30M) | Monthly total |
|---|---|---|---|
| Actual: $2 / $10 per MTok | $960.00 | $300.00 | $1,260.00 |
| Cancelled: $3 / $15 per MTok | $1,440.00 | $450.00 | $1,890.00 |
That's $630 a month, $7,560 a year, that this product isn't paying because Anthropic backed off the increase. Nobody sent an invoice for it, nobody wrote a change request — the bill just stayed where it was. If you budgeted for the September step-up like the original announcement told you to, you can put that line back.
Verified against Anthropic's own pricing docs, checked 2026-09-18. This is the complete current lineup plus the two previous-gen models still served for existing integrations.
| Modello | Ingresso | 5m cache write | 1h cache write | Cache hit | Output |
|---|---|---|---|---|---|
| Claude Fable 5 | $10.00 | $12.50 | $20.00 | $1.00 | $50.00 |
| Claude Opus 5 | $5.00 | $6.25 | $10.00 | $0.50 | $25.00 |
| Claude Opus 4.8 (same rate) | $5.00 | $6.25 | $10.00 | $0.50 | $25.00 |
| Claude Sonnet 5 | $2.00 | $2.50 | $4.00 | $0.20 | $10.00 |
| Claude Sonetto 4.6 (prior gen) | $3.00 | $3.75 | $6.00 | $0.30 | $15.00 |
| ClaudeHaiku4.5 | $1.00 | $1.25 | $2.00 | $0.10 | $5.00 |
All prices per million tokens. Worth noticing: Sonnet 5 at $2/$10 now undercuts its own previous generation, Sonnet 4.6 at $3/$15 — the new model is both better and 33% cheaper than the one it replaced, which almost never happens on a launch.
I ran that same 480M-input / 30M-output monthly workload through all four current models, no other changes. The spread surprised me less for the size of it than for how tidy it is.
| Modello | Costo mensile | vs Haiku 4.5 |
|---|---|---|
| ClaudeHaiku4.5 | $630.00 | 1.0x |
| Claude Sonnet 5 | $1,260.00 | 2.0x |
| Claude Opus 5 | $3,150.00 | 5.0x |
| Claude Fable 5 | $6,300.00 | 10.0x |
Sonnet costs exactly double Haiku, Opus exactly five times Haiku, Fable exactly ten. Because input and output are both round multiples of Haiku's rate across every tier, the ladder holds for any workload shape, not just this one — which makes back-of-envelope tier comparisons genuinely reliable instead of a rough guess. The real decision isn't "which is cheapest," it's whether the task in front of you needs Opus-or-Fable-grade reasoning, because the cost difference for getting that wrong compounds fast at volume.
Say you're running a coding-agent product that resends the same 6,000-token repo context on every follow-up turn inside a session, and a session averages 20 turns inside a 5-minute window — squarely what Anthropic's 5-minute cache is built for. On Sonnet 5: a 5-minute cache write costs $2.50/MTok (1.25x base input), a cache hit costs $0.20/MTok (0.1x base input).
| Per 20-turn session | At 50,000 sessions/mo | |
|---|---|---|
| No caching (120,000 ctx tokens × $2/MTok base) | $0.2400 | $12,000.00 |
| With caching (1 write + 19 hits) | $0.0378 | $1,890.00 |
That's $10,110 a month back, an 84% cut, for changing nothing about the product — just turning caching on for context that was already being resent verbatim. If your agent or RAG app isn't caching its repeated context yet, this is the single highest-leverage line item on the whole rate card.
For anything that doesn't need a live response — overnight classification, bulk tagging, backfills — the Batch API is a flat 50% off both input and output, on every model, no exceptions. A 100M-input / 100M-output classification run on Haiku 4.5 is $600 at standard rates and exactly $300 batched. It's the simplest saving on this whole page: no architecture change, just a different endpoint and an async result.
Default to Sonnet 5 unless you've specifically measured that a task needs more. At $2/$10 it now undercuts the model it replaced while scoring higher on Anthropic's own benchmarks, which is the rare case where "use the new default" is also the cheap option. Reach for Opus 5 only where the reasoning gap is visible in your own evals — not because it feels safer — and reserve Fable 5 for the genuinely hard, long-horizon runs where a 10x bill over Haiku is obviously worth it. And turn prompt caching on before you optimize anything else; in the example above it beat every model-tier decision on this page.
New to usage-based LLM pricing? Start with the guide gratuite sui costi API.
Deploy it yourself: DigitalOcean: credito gratuito di $ 200 ↗ · Hostinger VPS ↗
Pricing verified live against Anthropic's official Claude API pricing docs (platform.claude.com/docs/en/about-claude/pricing), checked 2026-09-18. Reference estimates using Anthropic's published self-serve rates — Bedrock, Google Cloud and Microsoft Foundry pricing, enterprise volume discounts and future rate changes vary; confirm current pricing at the source before budgeting. Workload figures (request volumes, token counts) are illustrative scenarios computed against real published per-token rates, not vendor-supplied numbers.