Claude Fable 5.1 Pricing: $10/$50, Cache Reads $0.25

Anthropic shipped Claude Fable 5.1 on 1 September 2026 as a point release on top of Claude Fable 5. The headline price didn't move, the context window didn't move, and the model ID barely moved — a combination that makes it easy to file the release under "minor" and move on.

It isn't minor. The most consequential change is buried in a pricing footnote, and there's a genuine breaking change that returns a 400 error on existing agent code. This guide covers both, with every number traced back to Anthropic's own documentation.

One thing has changed since Fable 5.1 shipped, and it changes the buying decision: Claude Opus 5.5 launched on 22 September 2026 at $4/$20 per million tokens. Claude Opus 5 — the model this guide originally pointed you to as the cheaper default — is now in Anthropic's legacy list. Every figure below has been re-checked against Anthropic's live docs and against Artificial Analysis's Intelligence Index v4.3.2 as of 5 October 2026.

What is Claude Fable 5.1?

Claude Fable 5.1 is the model Anthropic's docs point you to "for demanding reasoning and long-horizon agentic work," and the Choosing a model page still calls it "Anthropic's most capable model open to all customers." It is explicitly not the default recommendation. Since 22 September 2026 that sentence names a different model: the models overview now says to "start with Claude Opus 5.5 for most workloads" and reach for Fable 5.1 only "if your evals at xhigh or max effort still fall short on demanding reasoning or long-horizon agentic work." A sibling model, Claude Mythos 5.1, shares its exact specifications and pricing but is restricted to Anthropic's invitation-only Project Glasswing.

Worth stating plainly, because it is the single most useful fact on this page: Anthropic's own benchmark table in the Opus 5.5 announcement puts Opus 5.5 ahead of Fable 5.1 on all nine evals it published, and Artificial Analysis's Intelligence Index v4.3.2 agrees. So "most capable model open to all customers" is a positioning statement about long-horizon work, not a claim that Fable 5.1 wins on the scoreboard. It doesn't. The section on how Fable 5.1 compares with Opus 5.5 has the numbers.

Two details in that table matter more than they look. The 1M-token context window is billed at standard per-token rates across the whole window — a 900k-token request costs the same per token as a 9k-token one, with no long-context premium. And Fable 5.1 uses the tokenizer introduced with Claude Opus 4.7, which produces roughly 30% more tokens for the same text than the pre-4.7 tokenizer. If you are comparing costs against Claude Sonnet 4.6 or anything older, the token counts are not like-for-like.

What changed from Claude Fable 5?

Anthropic groups the changes into three breaking and five additive. The breaking ones are the reason you should not treat this as a drop-in string swap. For where Fable 5 started, see our Claude Fable 5 launch guide.

Forced tool use now returns an error

This is the one most likely to break production code. On Fable 5.1, tool_choice set to {"type": "any"} or {"type": "tool", "name": "..."} returns a 400 invalid_request_error:

tool_choice: type "tool" and "any" are not supported for this model.

tool_choice: {"type": "auto"} (the default) and {"type": "none"} still work. The same validation applies to the token-counting endpoint, so you can't even price a request that uses forced tool calls.

The reasoning is coherent: thinking is always on for this model, and a forced tool call would skip it — the model would write its working-out into the tool arguments instead, which degrades argument quality. Anthropic's recommended replacements are to keep tool_choice: {"type": "auto"} and set strict: true with strict tool use, move the schema to structured outputs, or simply instruct the model in the prompt ("Use the get_weather tool to answer"). The docs claim Fable 5.1 follows explicit tool instructions reliably.

Thinking blocks are bound to the model and to the conversation prefix

Every thinking block now records which model produced it, and preservation runs one way only. Fable 5.1 can read earlier models' thinking blocks; no earlier model can read Fable 5.1's. A router or fallback that switches models mid-conversation silently drops those blocks — they aren't billed and don't count toward input_tokens, but the reasoning is gone for turns that run elsewhere.

Separately, modifying anything before a Fable 5.1 thinking block — the system prompt, the tools array, or an earlier message — invalidates every thinking block after it and errors on the next request, enforced for accounts created on or after 31 August 2026. Treat the conversation as append-only: injecting a per-request reminder into history and deleting it next turn is exactly the pattern that now breaks. Claude Code, claude.ai, Claude Managed Agents and the Claude Agent SDK handle this for you; if your code builds the messages array by hand, this is the migration item to budget time for.

Cheaper cache reads, and three new beta features

The additive changes are the cache-read price cut (covered in detail below), plus three betas worth knowing about:

• Per-message effort (mid-conversation-output-config-2026-07-01) — change the effort level partway through a conversation without invalidating the prompt cache. Raise it for a hard step, drop it for routine ones.
• Turn-scoped system messages (mid-conversation-system-clear-at-2026-08-21) — set clear_at: "next_user_message" on a system message so it carries system-prompt authority for one turn, then stops rendering. It stays in messages, so history remains append-only and the cache keeps matching. A cleared message costs no input tokens.
• Readable progress updates (thinking-display-updates-2026-08-18) — set thinking.display to "updates" to receive between-tool-call status lines as text while raw reasoning stays hidden.

Anthropic also documents a set of behaviour changes that need no code change but will show up in your evals. Fable 5.1 batches parallel tool calls less consistently than Fable 5 (sometimes one call per turn where Fable 5 issued several), writes fewer progress updates, answers from memory more often at low effort instead of searching, writes denser prose with less formatting, is more likely to reproduce source passages in summaries without marking them as quotations, and is more likely to rewrite a whole file when a targeted edit would do. None of those reduce answer quality, but several cost tokens and wall-clock time.

What does Claude Fable 5.1 actually cost?

Here is the full price sheet, from Anthropic's pricing page.

The cache-read row is the whole story of the 5.1 release. Anthropic's pricing table carries a footnote saying cache hits and refreshes on Fable 5.1 and Mythos 5.1 are priced at 0.025x the base input price, against the standard 0.1x. Ten percent of $10 is the $1.00 Fable 5 charges; 2.5% of $10 is $0.25. That is a 75% reduction with no change to the sticker price, and because it lives in a multiplier footnote rather than a headline it is easy to miss entirely.

What is easier to miss now is the row next to it. Anthropic added a third cache multiplier with Opus 5.5 — 0.05x, per a second footnote on the same table — and 5% of $4 is $0.20/MTok. Fable 5.1 has the lowest cache multiplier in the lineup but no longer the lowest cache price. If you picked Fable 5.1 for cheap cache reads, that reason expired on 22 September.

A worked example on a real agent workload

Take a coding agent with a 200,000-token cached prefix (system prompt, repo map, project conventions, a few reference files), running a 40-turn session. Each turn re-reads the cached prefix, adds ~2,000 tokens of new input, and produces ~1,500 output tokens. Use the 1-hour cache so the prefix survives the session.

That is a 38% cut on identical work, purely from the cache-read line. It also matches what Anthropic claims in the announcement: roughly 25% less than Fable 5 for typical workloads, with savings "up to approximately 45%" on agentic work. The more cache-heavy your workload, the closer you sit to the top of that range.

The Opus 5.5 column shows the same session at $4.72 against Fable 5.1's $9.80. Read that as a rate comparison, not a forecast, because the table deliberately holds token counts identical across all three models — and they do not behave identically. Anthropic's own framing is "at default settings it will cost 40% less than Opus 5 on typical workloads," with cache reads "60% less than Opus 5."

The per-token price and the per-job bill are different numbers, and this is where most comparisons go wrong. Artificial Analysis publishes a cost-per-completed-Index-task figure that captures effort, verbosity and price together, and it tells a sharper story than any rate card. On v4.3.2: one Index task costs $7.63 on Claude Fable 5.1 (Max, Default Fallback), $5.98 on Claude Opus 5.5, $5.86 on Claude Opus 5 (Max) and $8.75 on Claude Fable 5 (Max).

Two conclusions, both counter-intuitive. Against Fable 5.1, Opus 5.5 is genuinely the better buy — 21.6% less per finished task and 4.27 Index points stronger — but the saving is a fifth, not the 60% the rate card implies. And against Opus 5, Opus 5.5 is 2% more expensive per task despite charging 20% less per token, because it spends materially more tokens reaching the same answer. If you are moving off Opus 5, move for the capability and the fact that Anthropic has reclassified it, not for a cost saving that does not survive measurement.

An honest caveat on Fable 5.1's own numbers: at $10 in and $50 out, this is still the most expensive model Anthropic sells, and the discount only applies to cache reads. Writes are unchanged, and the minimum cacheable prompt is 512 tokens. If your workload is short one-shot calls with no reusable prefix, Fable 5.1 costs you exactly what Fable 5 did — and 2.5x Opus 5.5's rate on both input and output.

How good is Claude Fable 5.1?

Worth separating what Anthropic reports from what third parties measure.

Vendor benchmarks

From Anthropic's own launch announcement:

Terminal-Bench-Science more than doubles, which is the single biggest jump on the board. The general-purpose numbers are more incremental: CursorBench moves 70.5% to 73.4%, and Humanity's Last Exam without tools moves 57.8% to 60.9%. Notably, Anthropic did not publish SWE-bench Verified, GPQA Diamond, ARC-AGI or tau-bench figures for this release, so anyone quoting those for Fable 5.1 is not quoting Anthropic.

Independent measurement

Artificial Analysis rebased its Intelligence Index from v4.1.1 to v4.3.2, and because the benchmark basket changed there is no conversion factor between the two — every model's score moved. On v4.3.2, captured 5 October 2026, and reading the variant AA's own model page displays:

Three things fall out of that table that no single headline score tells you.

Fable 5.1's lead is a Max-effort lead only. At Max, Fable 5.1 (53.35) is 2.57 points ahead of Opus 5 at Max (50.78). At each model's default effort, the ordering reverses: Opus 5 scores 49.68 against Fable 5.1's 46.82. If you compare an unqualified Fable 5.1 against another model's Max tier, you are comparing unlike things — always check which variant a score belongs to.

Opus 5.5 is the highest score in the v4.3.2 field at 57.62, 4.27 points clear of Fable 5.1 at Max, and it costs 21.6% less per completed Index task ($5.98 against $7.63). There is no effort tier at which Fable 5.1 catches it. Note that the per-task column, not the index-suite total, is the comparable cost figure: AA runs a different number of tasks per model — about 1,720 for Fable 5.1 at Max against roughly 1,460 for Opus 5.5 — so suite totals answer "what did this model's run cost" and nothing more.

Even Sonnet 5.5 at Max outscores Fable 5.1 — 56.00 against 53.35, from a $2/$10 model. Fable 5.1 is the most expensive model in Anthropic's catalogue and third on Anthropic's own index entry.

Two figures this guide used to quote are gone rather than restated: AA retired the standalone Coding Index and Agentic Index composites that existed under v4.1.1, so the old 81.6 / 61.3 pair has no v4.3.2 equivalent and converting it would be invented arithmetic. The old "first out of 196 models" rank is also dropped — AA's field has grown past 200 entries and Fable 5.1 is no longer top of it on either count.

The same evaluation surfaces the trade-off, and the variant split makes it vivid. At Max effort, Fable 5.1 streams 61.44 output tokens per second but takes a median 259 seconds to produce its first chunk. At default effort the same model starts answering in 3.4 seconds — and scores 6.5 Index points lower. That is what "adaptive thinking, always on" costs you: the intelligence and the wait are the same purchase. Anthropic's own comparison table simply lists Fable 5.1's latency as "Slower," the slowest in the current lineup, against "Moderate" for Opus 5.5. For a batch job or a background agent, irrelevant. For anything a human is watching, disqualifying.

How does Claude Fable 5.1 compare with Claude Opus 5.5?

Short version, and it is not the answer this guide gave in September: Opus 5.5 wins on capability, price and speed simultaneously. Fable 5.1's remaining case is positioning, not scoreboard.

Anthropic published a same-harness, three-way table in the Opus 5.5 announcement. Fable 5.1 loses every row.

Those are a different benchmark generation from the Fable 5.1 launch table higher up the page — CursorBench 4.0 not 3.2.0, OSWorld 2.1 not 2.0, GDPval-AA v2.1 not v2 — so don't read across between the two tables. Within this one, the margins run from 0.6 points on Chartography to 10.6 on Terminal-Bench 4.0.

Both models have a 1M-token context window, 128K max output, and always-on adaptive thinking that cannot be disabled. Both reject forced tool use with a 400 error — that is no longer a Fable-only constraint, which removes one of the reasons to stay on Opus. Where they diverge: Opus 5.5 is 60% cheaper on input and output per token (21.6% cheaper per completed AA Index task, once its heavier token appetite is counted), cheaper on cache reads ($0.20 vs $0.25), "Moderate" rather than "Slower" latency, generates output "more than 30% faster than Opus 5" per Anthropic, defaults to medium effort rather than high, supports 300K output tokens on the Batch API with the output-300k-2026-03-24 beta header, and supports Fast mode at $8/$40. Fable 5.1 has none of those.

Anthropic's escalation rule is still the right one, just re-pointed: run your evals on Opus 5.5 at xhigh or max effort first, and move to Fable 5.1 only if they still fall short on demanding reasoning or long-horizon agentic work. For the full head-to-head, see Claude Fable 5.1 vs Opus 5 and Opus 5.5, and our Claude Opus 5.5 complete guide for Opus 5.5's own specs and benchmarks.

Is Claude Fable 5.1 available on Claude Pro and Max, or API only?

Partly — and the fine print matters more than the yes. Every paid plan lists Opus, Sonnet and Haiku; Fable is the exception on every one of them. On Anthropic's plan comparison, Fable is not available on Free at all. On Pro ($20/month, or $17/month billed annually at $200 upfront) Fable is reachable only through separate usage credits, not your included allowance. On both Max tiers (from $100/month, in 5x and 20x usage variants) and on Team Standard and Premium, Fable carries a "50% of weekly limits" note — it burns your weekly allowance roughly twice as fast as the other models. Enterprise lists it as available.

In Claude Code, Fable is not the default on any plan; it must be explicitly selected, and depending on your plan and seat tier, Fable usage may bill to usage credits rather than your plan's included limits. If you switch a Pro or Max subscription to Fable 5.1 and leave it there, expect to hit limits sooner or start drawing down credits. Our guide to Claude Fable usage credits and plan limits covers how that accounting works.

One more constraint that matters for regulated buyers: Fable 5.1 and Mythos 5.1 carry 30-day data retention and are not available under zero data retention unless Anthropic expressly authorises it. Both are Covered Models, as Fable 5 and Mythos 5 were.

How do you use Claude Fable 5.1?

The model string is claude-fable-5-1 on the Claude API, Google Cloud, Microsoft Foundry and Claude Platform on AWS, and anthropic.claude-fable-5-1 on Amazon Bedrock. Migrating from Fable 5 is a one-line change plus the tool_choice audit described above.

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-fable-5-1",
    "max_tokens": 4096,
    "messages": [
      {"role": "user", "content": "Summarise this repo migration plan in three steps."}
    ]
  }'

In Python, with prompt caching switched on — which is the point of this release:

import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-fable-5-1",
    max_tokens=4096,
    system=[
        {
            "type": "text",
            "text": LARGE_SYSTEM_PROMPT,   # must be >= 512 tokens to be cacheable
            "cache_control": {"type": "ephemeral", "ttl": "1h"},
        }
    ],
    messages=[{"role": "user", "content": "What changed in the auth module this week?"}],
)

print(response.usage.cache_read_input_tokens)   # these bill at $0.25 / MTok
print(response.usage.cache_creation_input_tokens)

Three things to get right if you want the saving:

1. Meet the 512-token minimum. Anything shorter is not cacheable, and you pay full input price.
2. Pick the right TTL. The 5-minute write costs 1.25x base input, the 1-hour write costs 2x. At 0.025x reads, a 5-minute cache pays for itself after a single hit and a 1-hour cache after two. Both break even far faster than they did on Fable 5.
3. Keep history append-only. Editing an earlier turn invalidates thinking blocks and breaks the cache prefix. Use turn-scoped system messages for per-turn reminders and server-side context editing or compaction for trimming, not client-side surgery on messages.

In Claude Code, Fable 5.1 is the model behind the fable alias. Select it with /model fable, claude --model fable, claude --model claude-fable-5-1, or export ANTHROPIC_MODEL=fable. Claude Code v2.1.255 or later is required for Fable 5.1. If you're weighing this against other agent harnesses, our complete guide to AI coding agents covers the wider landscape.

Should you upgrade from Claude Fable 5?

If you already run Fable 5 with prompt caching, upgrade — the same money buys the same tokens at higher benchmark scores, and cache-heavy sessions get materially cheaper. Budget an afternoon for the migration: strip any tool_choice of type any or tool, verify your code never edits earlier turns, then re-run your evals and re-tune effort from the high default.

If you run Fable 5 without caching, the upgrade is free but the cost story doesn't apply to you. Fix the caching first; that's where the money is, on either model.

And if you are not on Fable at all, the honest answer is that Anthropic doesn't think you should be — its documentation points you at Opus 5.5, which is 60% cheaper per token, 21.6% cheaper per completed Artificial Analysis Index task, faster to first token, and ahead of Fable 5.1 on all nine benchmarks Anthropic published in September. The decision rule, in one line: run Opus 5.5 at xhigh or max effort, measure, and only reach for Fable 5.1 if your evals still fall short on hours-long autonomous work. Fable 5.1 remains the right tool for exactly that case, where a minutes-long first token is acceptable and correctness is worth 2.5x the token rate. It is the wrong tool for anything interactive, and as of 22 September it is no longer the default answer to "which Claude is best."

FAQ

What is Claude Fable 5.1?

Claude Fable 5.1 is Anthropic's most capable Claude model, released on 1 September 2026 as a successor to Claude Fable 5. It has a 1M-token context window, 128K max output, always-on adaptive thinking, and a June 2026 knowledge cutoff. Anthropic positions it for demanding reasoning and long-horizon agentic work rather than as a general default.

How much does Claude Fable 5.1 cost?

$10 per million input tokens and $50 per million output tokens on the Claude API — identical to Fable 5. Cache writes are $12.50/MTok for 5-minute and $20/MTok for 1-hour. Cache reads are $0.25/MTok. The Batch API halves input and output to $5 and $25 per million tokens.

What's new in Fable 5.1 versus Fable 5?

Cache reads dropped 75% to $0.25 per million tokens, and benchmarks improved across agentic coding, research, vision and computer use. Three changes are breaking: forced tool use now errors, thinking blocks are bound to the producing model, and editing earlier turns invalidates them. Per-message effort, turn-scoped system messages and readable progress updates are new betas.

Is Fable 5.1 better than Opus 5 or Opus 5.5?

Better than Opus 5, worse than Opus 5.5. Against Opus 5 it leads every benchmark Anthropic published and scores 53.35 to 50.78 on Artificial Analysis's Intelligence Index v4.3.2 (Max variants, October 2026) — though at each model's default effort Opus 5 actually wins, 49.68 to 46.82. Against Opus 5.5 it loses all nine benchmarks in Anthropic's own announcement table and trails 53.35 to 57.62 on the same index, while costing 2.5x as much per token. Opus 5 is now a legacy model; Opus 5.5 is the default.

Is Claude Fable 5.1 available on the Pro plan?

Only through usage credits, not your included Pro allowance. Anthropic's plan comparison marks Fable as "usage credits" on Pro, "50% of weekly limits" on both Max tiers and on Team Standard and Premium, available on Enterprise, and not available at all on Free. Fable is never the default model on any plan — you have to select it explicitly, and in Claude Code it may draw down credits rather than included limits depending on your plan and seat tier.

What is Fable 5.1's context window?

1,000,000 tokens, which is both the default and the maximum. The entire window is billed at standard per-token rates with no long-context premium. Maximum output is 128,000 tokens; unlike Opus 5.5, Opus 5, Sonnet 5.5 and Sonnet 5, Fable 5.1 does not support the 300K-output batch beta with the output-300k-2026-03-24 header. Roughly 555k English words fit in 1M tokens on the current tokenizer.

Is Claude Fable 5.1 cheaper than Claude Opus 5.5?

No, on every line item. Fable 5.1 is $10/$50 per million input/output tokens against Opus 5.5's $4/$20; cache reads are $0.25 against $0.20; 1-hour cache writes $20 against $8; batch $5/$25 against $2/$10. The honest measure of total cost is Artificial Analysis's per-task figure, which counts verbosity as well as rates: $7.63 per Index task on Fable 5.1 (Max) against $5.98 on Opus 5.5 on v4.3.2 — a 21.6% saving, smaller than the 60% rate gap implies. Fable 5.1 has the lowest cache-read multiplier in the lineup (0.025x) but not the lowest cache-read price.

Should I upgrade from Fable 5?

Yes if you use prompt caching, since identical workloads get 25-45% cheaper at the same headline price. Before switching, remove any tool_choice of type any or tool (it now returns a 400 error), make sure your code never edits earlier conversation turns, and re-run your evals — several default behaviours changed.

Is Claude Fable 5 deprecated?

Not deprecated, but reclassified. Anthropic's models overview lists Claude Fable 5 under "Legacy models (still available)" — a list that as of October 2026 also contains Claude Opus 5, Opus 4.8, Opus 4.7, Opus 4.6, Opus 4.5, Sonnet 5 and Sonnet 4.6. Fable 5 continues to serve requests at $10/$50 with $1/MTok cache reads. Fable 5.1 is still marked "Active (latest)" with a retirement commitment of not sooner than 1 September 2027.