Claude Fable 5 benchmarks

Claude Fable 5 is a proprietary model from Anthropic, released 9 Jun 2026. It ranks 2nd of 51 on our overall leaderboard and 5th of 51 for reasoning. At A$28.08 per million tokens (blended), it costs 7.8× the median model we track. It can be called from AWS Bedrock in Sydney and Melbourne and Google Vertex AI in Sydney, but only through global routing, so requests may be processed outside Australia.

#2 OF 51 OVERALLA$28.08 / 1M TOKENS1M CONTEXTUPDATED 23 SEP 2026
CONTEXT WINDOW
1,000,000 tokens
MAX OUTPUT
128,000 tokens
INPUT
text, image
WEIGHTS
Proprietary
RELEASED
9 Jun 2026
ANTHROPIC DOCS
§ 02 — RESULTS

Benchmark results.

Independent test scores are for the adaptive reasoning, max effort, opus 4.8 fallback setting. “Among tracked” ranks Claude Fable 5 against the 51 models on this site; LMArena ranks run across its full leaderboard.

ALL RESULTS7 OF 8 CORE SIGNALS
Claude Fable 5 benchmark results
BENCHMARKRESULTAMONG TRACKEDSOURCE DETAIL
INDEPENDENT TESTS · ARTIFICIAL ANALYSIS
AA Intelligence Index49.6#5 of 51Adaptive reasoning, max effort, opus 4.8 fallback
AA Coding Index76.5#7 of 42Adaptive reasoning, max effort, opus 4.8 fallback
GPQA Diamond92.6%#17 of 44Adaptive reasoning, max effort, opus 4.8 fallback
Humanity’s Last Exam55.5%#3 of 51Adaptive reasoning, max effort, opus 4.8 fallback
AA-LCR82.3%#15 of 51Adaptive reasoning, max effort, opus 4.8 fallback
Terminal-Bench 2.184.6%#12 of 42Adaptive reasoning, max effort, opus 4.8 fallback
τ²-Bench (banking)38.1%#20 of 41Adaptive reasoning, max effort, opus 4.8 fallback
SciCode61.0%#3 of 46Adaptive reasoning, max effort, opus 4.8 fallback
BLIND HUMAN VOTES · LMARENA
LMArena Text30,057 votes · ±51506#1 of 42#1 of 402 on LMArena · 13 Sep 2026
LMArena WebDev11,824 votes · ±71627#9 of 45#12 of 129 on LMArena · 22 Sep 2026
LMArena Vision11,304 votes · ±81310#1 of 33#1 of 152 on LMArena · 13 Sep 2026
LMArena Document8,497 votes · ±81499#1 of 26#1 of 44 on LMArena · 13 Sep 2026
LMArena Search41,795 votes · ±81229#2 of 12#2 of 34 on LMArena · 24 Aug 2026
LMArena Agent38,293 sessions#5#4 of 37#5 of 46 on LMArena · 15 Sep 2026
§ 03 — COST

What Claude Fable 5 costs in Australian dollars.

List prices converted at the snapshot’s RBA rate, excluding GST. Reasoning models are billed for their thinking as output tokens, so heavier settings cost more than the job estimates show.

LIST PRICE
Claude Fable 5 price per million tokens
PER MILLION TOKENSAUDUSD LIST
Input tokensA$14.04US$10.00
Output tokensA$70.19US$50.00
Blended (3 in : 1 out)A$28.08US$20.00
EVERYDAY JOBSESTIMATES
Estimated cost of common jobs in Australian dollars
JOBCOST (AUD)MEDIAN MODEL
Customer support replyPER 1,000 REPLIESA$49.14A$6.58
Summarise a 30-page documentPER 100 DOCUMENTSA$33.69A$4.74
Agentic coding taskPER 10 TASKSA$26.67A$3.72
§ 04 — AUSTRALIA

Running Claude Fable 5 in Australia.

It can be called from AWS Bedrock in Sydney and Melbourne and Google Vertex AI in Sydney, but only through global routing, so requests may be processed outside Australia. How the platforms compare →

AUSTRALIAN CLOUD REGIONS
Claude Fable 5 availability in Australian cloud regions
PLATFORMAVAILABILITYDETAIL
AWS Bedrock · SydneyGLOBAL

Provider docs, checked 23 Sep 2026

AWS Bedrock · MelbourneGLOBAL

Provider docs, checked 23 Sep 2026

Azure · Australia EastNOT IN AU

Provider docs, checked 23 Sep 2026

Google Vertex AI · SydneyGLOBAL

Provider docs, checked 23 Sep 2026

§ 06 — QUESTIONS

Claude Fable 5, answered.

Claude Fable 5 lists at US$10.00 per million input tokens and US$50.00 per million output tokens — A$14.04 and A$70.19 at A$1 = US$0.7123 (RBA, 22 Sep 2026), excluding GST. A typical customer support reply works out at about A$49.14 per 1,000 replies, before any reasoning tokens.

Claude Fable 5 can be called from AWS Bedrock in Sydney and Melbourne and Google Vertex AI in Sydney, but only through global routing, so requests may be processed outside Australia. Availability comes from the cloud providers’ own documentation; check your provider’s current terms before relying on it for data-residency obligations.

It ranks 8th of 42 on our coding ranking, with an Artificial Analysis Coding Index of 76.5 and an LMArena WebDev rating of 1627 (12th of 129). The current leader is Claude Fable 5.1.

It ranks 12th of 41 on our agents ranking, completing 38.1% of τ²-Bench banking customer-service tasks and 84.6% of Terminal-Bench 2.1 tasks.

Anthropic documents a context window of 1,000,000 tokens (1M), with up to 128,000 output tokens. On AA-LCR, which tests reasoning across ~100,000-token document sets, it scores 82.3%.

PUT THE COMPARISON TO WORK

Need help choosing and using AI for your business?