Skip to content
Tier A — Frontier
Runs in:USMade in:United States

Archived

This model has been discontinued by the provider. Historical data is preserved.

No longer available since June 14, 2026.

Anthropic

Claude Fable 5

Tier A — Frontier · 1M tokens

Tokonomix Editorial Team·Reviewed by Mes Kalkan·

Claude Fable 5 is Anthropic's vision-and-reasoning model with a one-million-token context variant. In our own image-QC pilot it was the steadiest pair of eyes we tested — so we made it the default vision proposer in our image-consensus panel. Here is what we measured, and where we did not.

The property we wanted from a vision proposer was not cleverness but steadiness — and on our pilot set, Fable 5 was the steadiest.

Tokonomix vision-QC pilot, 9 June 2026
Section 01

Pricing history

Direct provider rates per million tokens, plus a typical-conversation cost estimate.

💰
API rates — Claude Fable 5
$10.00 per 1M input tokens
$50.00 per 1M output tokens
≈ $0.0160 per typical conversation (800 tokens)
Input vs output price (per 1M tokens)
per 1M input tokens$10.00
per 1M output tokens$50.00

Pricing over time

Input & output per 1M tokens · step-line = price changes

$10.00

input / 1M

— stable

$50.00

output / 1M

— stable

2026-06-102026-06-142026-06-14
Input
Output
Price change
⟳ synced weekly
Section 02

Strengths & weaknesses

Drawn from benchmark results and aggregated community feedback on real use-cases.

Strengths

Most stable vision proposer in our pilot (88% run-identical)Low false-alarm rate (7.1% on the 300-image baseline)Caught blind spots other panel models missed1M-token context variant for long, image-heavy material

Weaknesses

Solo recall 66.9% — a panel is needed to reach 87.5%Reasoning/coding numbers are early (single run, n=1)
Section 03

Capabilities

toolssource: manualvisionjson modepdf inputreasoningjson schemamodel class: mythosprompt cachingmax output tokens: 64000
Section 04

Frequently asked questions

On our 9 June 2026 pilot it was run-identical 88% of the time, flipped its answer only 3.9% of the time, and produced zero false positives — the steadiness you want at the front of a QC pipeline. It was a deliberate product decision based on that pilot.

We would rather hand you the texture — small pilot, one-day baseline, early reasoning numbers — than a clean story that does not survive the next run.

Tokonomix editorial
Section 05

Availability

Availability

No measurements yet

We haven't recorded enough API calls to show availability stats for this model. Data appears once the model starts receiving live traffic.

Image quality control pilot (2026-06-10)

Recall

66.9%

n=300

False-alarm rate

7.1%

n=300

Section 06

Tokonomix benchmark verdicts

⚖️
Endorsed by 2 judges
Independent LLM judges evaluated this model on our weekly intelligence tests
cohere/command-a100/100 · 1 runs
1 correct0 partial0 wrong100% accuracy
claude-sonnet-4-591/100 · 25 runs
22 correct1 partial2 wrong88% accuracy
2026-06-14

Claude Fable 5 shows improved coding, but reasoning and factual decline

Claude Fable 5 demonstrates a mixed performance shift from the previous window. The overall quality score decreased from 88.4 to 85.7, driven primarily by notable declines in reasoning and factual accuracy. Reasoning performance dropped significantly from a perfect 100 to 83, representing the most substantial category decline. Factual accuracy also fell from 68 to 60, continuing weakness in this area. However, coding performance improved from 97 to 98, maintaining strength in technical tasks. Latency showed marginal improvement, with p50 decreasing from 8318ms to 7986ms, though this remains relatively slow for real-time applications. The test sample size increased from 5 to 8 runs, providing somewhat greater statistical confidence. No new capabilities were detected in this window, suggesting a focus on performance tuning rather than feature expansion. Users should be aware that while coding tasks remain highly reliable, applications requiring strong reasoning or factual retrieval may see degraded results compared to the previous benchmark period. The overall trend indicates some regression in core capabilities that may warrant attention from development teams relying on this model for production workloads.

Quality

Latency p50

Test runs

0

Reasoning dropped from 100 to 83 Factual accuracy declined to 60 Coding improved slightly to 98 Latency improved to 7986ms
Section 07

Full model profile

Claude Fable 5: the steadiest pair of eyes we tested

Anthropic's Claude Fable 5 is a vision-and-reasoning model, and the variant we lean on most — claude-fable-5[1m] — carries a one-million-token context window. We are not going to recite a spec sheet. What we can offer instead is something most write-ups cannot: numbers from our own measurements, with the dates and sample sizes attached, and an honest note about where the evidence runs thin.

Why we made it our default vision proposer

On 9 June 2026 we ran a vision-QC pilot — a set of images, each with a known media-quality flaw to catch, fed to several vision models so we could compare how they behaved. Fable 5 stood out for one quality that matters more than raw cleverness: it was stable.

Across repeated runs on the pilot set it was run-identical 88% of the time, with a flip-rate of 3.9% — meaning it rarely changed its mind about the same image from one pass to the next. On that pilot set it produced zero false positives: it did not invent defects that were not there. It also caught blind spots that other models in the panel missed.

Stability is the property you want from a proposer. A model that flags a real defect today and shrugs at the identical image tomorrow is hard to build a process on. One that answers the same way every time — and does not cry wolf — is one you can put at the front of a pipeline. So we did: Fable 5 is the default vision proposer in our image-consensus panel. (That was a deliberate product decision, made on the strength of the pilot.) It is not in our text-judge pools — its job here is looking at images, not scoring text.

Today's baseline: 300 images, measured live

A pilot is a small thing. So we put the model through a larger run and we publish the result, live, on our vision-QC benchmark.

On 10 June 2026, against the mediaqc-v3-2026-06-10 dataset of 300 images, Fable 5 solo scored:

  • 66.9% recall — the share of real defects it caught. That tied for the best single-model result in the run.
  • 7.1% false-alarm rate — how often it flagged a clean image. For comparison, another strong vision model on the same run sat more than twice as high on false alarms, which is exactly the kind of difference that decides whether a QC step saves time or wastes it.
  • 60.3% class-matched — cases where it not only caught the defect but named the right category of problem.

Recall of two-thirds, solo, is useful but not a finished story — which is the point of running a panel rather than a single model. With Fable 5 sitting in the consensus panel, recall rose to 87.5%. The proposer's steadiness gives the council something reliable to build on; the council closes the gap that one model alone leaves open.

Reasoning and coding: early, honest signal

Vision is where we have the most evidence. On general capability we have less, and we will say so plainly.

In an internal evaluation on 9 June 2026 — a 682-test harness we run across models — Fable 5's results separated sharply from the field: it scored 91% and 73% on the two halves of that harness where the models we compared against scored 3% and 0% on the identical tasks. That is a large gap, and we treat it as a strong signal rather than a settled verdict, because a single harness can flatter or punish a model in ways a second one would correct.

Our first intelligence run on the platform, today, returned a reasoning score of 100 and a coding score of 97. Those are preliminary — a single run, n=1. We are reporting them because hiding early numbers is its own kind of dishonesty, but one run is one run. Watch the leaderboard as the sample grows; that is where these figures will firm up or move.

What the 1M-context variant is for

The claude-fable-5[1m] variant exists for the cases where context is the bottleneck rather than reasoning. A window this wide lets you hold a long document set, a large body of reference material, or an extended interaction in front of the model at once instead of chunking it and hoping nothing important falls out of view. Paired with vision, it suits work where you need the model to reason over a lot of material and look at the images embedded in it — long reports, document sets with figures, screenshots in context.

How to read these numbers

Everything above is measured, dated, and bounded by a sample size we have named. The vision-QC pilot was small; the 300-image baseline is larger but is one dataset on one day; the reasoning and coding figures are early. We would rather hand you that texture than a clean story, because the clean story is usually the one that does not survive the next run.

Claude Fable 5 is available through the Tokonomix gateway and catalogue, and you can put it in front of your own images on our vision-QC benchmark or watch it across tasks on the leaderboard.

Last automated test
Jun 14, 2026 · 04:58 UTC · Benchmark
P50 latency
P95 latency
Errors
1 / 6 runs
Last reviewed by Tokonomix Team·June 10, 2026