What are Claude Fable 5.2's benchmark scores?

As of 9 Nov 2026, Claude Fable 5.2 (max effort) scores 69.4 ±5.8 on the BenchLeader Index, top 10, strongest in knowledge (82) and weakest in long context (66). That rests on independent benchmark results only, and every score on this page links to its publisher.

How does Claude Fable 5.2 rank?

Claude Fable 5.2 (max effort) is #1 of 599 ranked configurations on 9 Nov 2026 (top 10). 35 other configurations score within its margin, so the exact position matters less than the band.

Is Claude Fable 5.2 better than Claude Opus 5.5?

As of 9 Nov 2026, Claude Fable 5.2 (max) and Claude Opus 5.5 (max) are neck and neck on quality (69.4 vs 68.9 on the BenchLeader Index, inside the ±7.4 margin of error), and Claude Opus 5.5 (max) costs 2.0× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.

When was Claude Fable 5.2 released?

Anthropic released Claude Fable 5.2 on 9 Nov 2026. BenchLeader has results for it at 6 reasoning-effort settings.

AnthropicReasoning modelNewReleased 9 Nov 2026Compare with another model

Claude Fable 5.2

Reasoning effort

Quality#1 of 335

69.4Index

Top 10%

+19.4 on the median

Price#239 of 251

$16.00per 1M tokens

Dearest quarter

19× the median

Speed#117 of 217

72tokens/s

Below the median

1.0× the median

No cheaper ranked model scores higher: it is on the price frontier.

As of 9 Nov 2026, Claude Fable 5.2 by Anthropic scores 69.4 on the BenchLeader Index (top 10 of 599 ranked configurations), with its best results in knowledge (82) and weakest in long context (66). API pricing: $8.00 per 1M input tokens and $40.00 per 1M output tokens ($16.00 blended at 3:1), list price on 9 Nov 2026. It streams about 72 tokens/s (OpenRouter). Released 9 Nov 2026.

Key facts (as of 9 Nov 2026)

Claude Fable 5.2 API pricing: $8.00 per 1M input tokens and $40.00 per 1M output tokens ($16.00 blended at 3:1) (list price, 9 Nov 2026).
Claude Fable 5.2 streams about 72 tokens/s (OpenRouter, 9 Nov 2026).
Claude Fable 5.2 was released on 9 Nov 2026 (Anthropic).

In detail. Claude Fable 5.2 is an Anthropic proprietary reasoning model, released 9 Nov 2026. Its best configuration (max reasoning effort) ranks #1 of 599 on the BenchLeader Index at 69.4 ±5.8, in the top ten. The ± is the point: 35 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in knowledge (82) and lowest in long context (66). At $16.00 per million tokens blended it is among the most expensive ranked models. Output speed of 72 tokens per second puts it slower than most, with a first token in 3.0 s. It has been measured at 6 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 9 Nov 2026.

Blended price: $16.00/M; $8.00 in · $40.00 out; via Artificial Analysis
Output speed: 72 tok/s; OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
First answer: 3.00 s
Context: 1M; via models.dev
Full answer: –
Cost per task: $26.54; average task, AA Index
Tokens per task: 274k; 70% of it reasoning
Released: 9 Nov 2026

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index69
Reasoning78
Coding80
Knowledge82
Long context66

Versions

Anthropic has shipped 3 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.

Model	Released	Index	Rank
NewClaude Fable 5.2maxthis page	9 Nov 2026	69.4	#1
Claude Fable 5.1max	1 Sep 2026	66.9	#9
Claude Fable 5	9 Jun 2026	66.6	#10

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.

Effort	Index	Rank	Speed	First answer	Chat reply cost	Categories
low	65.3	#20	72 tok/s	3.00 s	$0.015	Coding 69 · Knowledge 80 · Long context 65 · Reasoning 71
medium	66.3	#12	72 tok/s	3.00 s	$0.015	Coding 72 · Knowledge 80 · Long context 65 · Reasoning 73
high	68.9	#3	72 tok/s	3.00 s	$0.015	Coding 81 · Knowledge 80 · Long context 66 · Reasoning 77
xhigh	68.7	#4	72 tok/s	3.00 s	$0.015	Coding 79 · Knowledge 81 · Long context 66 · Reasoning 77
maxbest	69.4	#1	72 tok/s	3.00 s	$0.015	Coding 80 · Knowledge 82 · Long context 66 · Reasoning 78
not stated	–	–	72 tok/s	3.00 s	$0.015	Composite 73

What more thinking costs

Turning the effort up buys index points and multiplies the bill. Here max costs 8.6× low for +4.1 on the index — 1.3 points per doubling of spend. Compare that with other models.

Each point is one reasoning setting: its cost per Artificial Analysis index task (log scale) and its BenchLeader Index.

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmark	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0 Reasoningnot in indexeffort not stated	–	–	–	–	–	88.8%#15	LiveBench
Humanity's Last Exam (AA)not in index	59.7%#6	61.9%#4	66.2%#3	68.9%#2	74.3%#1	–	Artificial Analysis
HLE-Diamond (AA)	60.8%#12	66.3%#7	73.8%#3	75.8%#2	80.2%#1	–	Artificial Analysis
ARC-AGI-1	94.5%#44	97.5%#13	97.5%#13	98.5%#1	97.5%#13	–	ARC Prize
ARC-AGI-2	86.7%#30	89.6%#20	90.4%#15	93.8%#3	92.5%#6	–	ARC Prize
ARC-AGI-3	92.3%#17	96.8%#11	97.7%#6	97.3%#8	94.9%#16	–	ARC Prize
CritPtnot in index	29.4%#24	30.9%#15	32.9%#3	35.7%#1	35.4%#2	–	Artificial Analysis
CritPt v2 hard subset (AA)	60.0%#25	61.5%#22	73.0%#3	73.5%#2	76.5%#1	–	Artificial Analysis
MysteryMechanismnot in indexeffort not stated	–	–	–	–	–	51.8%#2	Vals AI

Coding

Benchmark	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0 Codingnot in indexeffort not stated	–	–	–	–	–	79.2%#11	LiveBench
SciCode (AA)not in index	57.8%#29	59.7%#16	59.5%#18	65.6%#3	67.7%#1	–	Artificial Analysis
Code Migration v1.1not in indexeffort not stated	–	–	–	–	–	83.3%#2	Vals AI
ProgramBenchnot in indexeffort not stated	–	–	–	–	–	19.0%#1	Vals AI
SWE-bench Multimodalnot in indexeffort not stated	–	–	–	–	–	62.3%#2	Vals AI
Vibe Code Bench v1.1not in indexeffort not stated	–	–	–	–	–	93.8%#1	Vals AI
Terminal-Bench 4.0 (AA)not in index	54.5%#21	55.0%#19	59.6%#7	61.6%#5	59.6%#7	–	Artificial Analysis
FrontierCode 1.1 (AA)not in index	60.6%#2	64.0%#1	55.9%#5	55.5%#6	56.2%#4	–	Artificial Analysis
FrontierCode 2.0 (AA)	38.4%#16	41.3%#10	49.1%#1	47.6%#3	48.8%#2	–	Artificial Analysis

Agents & tools

Benchmark	low	medium	high	xhigh	max	not stated	Source
LiveBench Long Agenticnot in indexeffort not stated	–	–	–	–	–	82.9%#1	LiveBench
LiveBench Browsernot in indexeffort not stated	–	–	–	–	–	86.9%#7	LiveBench
GDPval-AA v2.1not in index	1546#37	1591#31	1718#13	1751#8	1831#6	–	Artificial Analysis
Finance Agent v2.1not in indexeffort not stated	–	–	–	–	–	69.5%#1	Vals AI
Harvey's Legal Agent Benchmark v1.1not in indexeffort not stated	–	–	–	–	–	5.0%#28	Vals AI
Legal Research Benchnot in indexeffort not stated	–	–	–	–	–	46.1%#19	Vals AI
Tax Agent Benchnot in indexeffort not stated	–	–	–	–	–	32.1%#26	Vals AI
Terminal-Bench 4.0 (Vals)not in indexeffort not stated	–	–	–	–	–	65.2%#1	Vals AI
Terminal-Bench Sciencenot in indexeffort not stated	–	–	–	–	–	55.7%#2	Vals AI
AutomationBenchnot in index	68.0%#18	68.8%#13	71.4%#9	74.2%#5	76.4%#4	–	Artificial Analysis
GDP.pdf (AA)not in index	26.8%#26	23.8%#53	21.8%#68	21.4%#72	24.2%#47	–	Artificial Analysis
Harvey LAB-AA v1.1not in index	4.2%#12	6.4%#8	4.7%#11	3.9%#14	7.2%#6	–	Harvey
AA-Briefcase v1.1not in index	1563#29	1622#17	1677#9	1738#6	1785#3	–	Artificial Analysis
Terminal-Bench-Science 0.1 (AA)not in index	53.8%#15	60.5%#9	70.0%#2	67.6%#3	73.3%#1	–	Artificial Analysis

Knowledge

Benchmark	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0 Data Analysisnot in indexeffort not stated	–	–	–	–	–	81.5%#6	LiveBench
LiveBench Researchnot in indexeffort not stated	–	–	–	–	–	94.5%#1	LiveBench
AA-Omniscience v1.1	62.9#5	63.5#4	64.0#3	64.8#2	66.8#1	–	Artificial Analysis
BioMysteryBenchnot in indexeffort not stated	–	–	–	–	–	83.7%#1	Vals AI
Excel Modeling Benchmark v1.1not in indexeffort not stated	–	–	–	–	–	77.4%#1	Vals AI
AA-Omniscience v1.1: accuracynot in index	76.3%#5	77.0%#4	77.2%#3	78.0%#2	79.1%#1	–	Artificial Analysis
AA-Omniscience v1.1: non-hallucinationnot in index	43.6%#77	41.5%#79	42.0%#78	39.6%#84	40.7%#80	–	Artificial Analysis
HealthBench Professional (AA)not in index	66.9%#5	67.0%#4	67.6%#3	68.1%#1	67.7%#2	–	OpenAI

Multimodal

Benchmark	max	not stated	Source
LiveBench Designnot in indexeffort not stated	–	85.1%#1	LiveBench
AA-3Dnot in index	1271#1	–	Artificial Analysis

Long context

Benchmark	low	medium	high	xhigh	max	Source
AA-LCR v1.1	80.0%#88	80.0%#88	83.3%#22	82.7%#40	83.3%#22	Artificial Analysis

Composite

Benchmark	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0effort not stated	–	–	–	–	–	84.9%#1	LiveBench
AA Intelligence Index v5.0not in index	53.9#15	56.4#9	61.8#3	62.2#2	64.6#1	–	Artificial Analysis
Vals Indexnot in indexeffort not stated	–	–	–	–	–	71.3#1	Vals AI
Vals Finance Indexnot in indexeffort not stated	–	–	–	–	–	63.6%#6	Vals AI
Vals Science Indexnot in indexeffort not stated	–	–	–	–	–	63.7%#2	Vals AI

Where it wins

Benchmarks where this configuration ranks in the top five of every configuration measured.

API pricing: where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

Provider	Speed	First token	Input $/M	Output $/M	Context	Quantisation
Anthropic	72 tok/s	3.00 s	$8.00	$40.00	1M	–
Azure	64 tok/s	4.24 s	$8.00	$40.00	1M	–

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.200 per 1M. Reasoning tokens are not modelled.

Workload	Tokens in / out	Cost	With caching	Time
Chat reply	400 / 300	$0.015	$0.013	7.2 s
Summarize a 30-page report	12,000 / 600	$0.120	$0.050	11.3 s
Code edit	6,000 / 1,500	$0.108	$0.073	23.8 s
Agentic coding session	60,000 / 4,000	$0.640	$0.289	58.6 s
Structured extraction	2,000 / 200	$0.024	$0.012	5.8 s

New this month

Compare Claude Fable 5.2 with…

Embed or cite

Show where Claude Fable 5.2 ranks on a model card or README. The badge redraws itself with every daily update, carries the data date and links back to this page. To see ranks like this on OpenRouter and Hugging Face model pages, get the Chrome extension.

[![Claude Fable 5.2 on BenchLeader: #1 overall](https://www.benchleader.com/badge/claude-fable-5-2.svg)](https://www.benchleader.com/models/claude-fable-5-2)

Cite this

The ranking changes daily, so a reference carries the data date. APA style:

BenchLeader. (2026, November 9). Claude Fable 5.2 on the BenchLeader Index (method v1.4) [Data set]. Retrieved November 9, 2026, from https://www.benchleader.com/models/claude-fable-5-2

BibTeX:

@misc{benchleader_claude_fable_5_2_20261109,
  author       = {{BenchLeader}},
  title        = {{Claude Fable 5.2 on the BenchLeader Index (method v1.4)}},
  year         = {2026},
  month        = nov,
  howpublished = {\url{https://www.benchleader.com/models/claude-fable-5-2}},
  note         = {Data as of 2026-11-09}
}

Every badge style, the embeddable top-10 widget and how to cite the Index

Questions people ask

What are Claude Fable 5.2's benchmark scores?: As of 9 Nov 2026, Claude Fable 5.2 (max effort) scores 69.4 ±5.8 on the BenchLeader Index, top 10, strongest in knowledge (82) and weakest in long context (66). That rests on independent benchmark results only, and every score on this page links to its publisher.
How does Claude Fable 5.2 rank?: Claude Fable 5.2 (max effort) is #1 of 599 ranked configurations on 9 Nov 2026 (top 10). 35 other configurations score within its margin, so the exact position matters less than the band.
How much does Claude Fable 5.2 cost per 1M tokens?: As of 9 Nov 2026, Claude Fable 5.2 costs $8.00 per 1M input tokens and $40.00 per 1M output tokens, $16.00 blended (three input tokens per output token).
Is Claude Fable 5.2 better than Claude Opus 5.5?: As of 9 Nov 2026, Claude Fable 5.2 (max) and Claude Opus 5.5 (max) are neck and neck on quality (69.4 vs 68.9 on the BenchLeader Index, inside the ±7.4 margin of error), and Claude Opus 5.5 (max) costs 2.0× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.
When was Claude Fable 5.2 released?: Anthropic released Claude Fable 5.2 on 9 Nov 2026. BenchLeader has results for it at 6 reasoning-effort settings.

Data as of 9 Nov 2026. Compare these configurations.

Cite as: BenchLeader, “Claude Fable 5.2: benchmarks, pricing, speed and rank”, data as of 9 Nov 2026, https://www.benchleader.com/models/claude-fable-5-2.