AnthropicReasoning modelClaude Fable 5
Claude Fable 5 is an Anthropic proprietary reasoning model, released 9 Jun 2026. Its best configuration (thinking reasoning effort) ranks #13 of 756 on the BenchLeader Index at 69.2 ±6.9, in the top 25. The ± is the point: 84 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 reasoning (87) and lowest in instruction following (63). At $20.00 per million tokens blended it is among the most expensive ranked models. Output speed of 66 tokens per second puts it slower than most, with a first answer in 97.2 s. It has been measured at 7 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 9 Jun 2026.
- Blended price
- $20.00/M
- $10.00 in · $50.00 out
- Output speed
- 66 tok/s
- measured by Artificial Analysis
- First answer
- 97 s
- first token 3.75 s
- Context
- 1M
- Full answer
- 105 s
- median, reasoning included
- Cost per run
- $8.75
- one full Intelligence Index run
- Released
- 9 Jun 2026
- Overall index69
- Reasoning87
- Agents & tools74
- Knowledge81
- Instruction following63
- Long context67
- Composite86
Versions
Anthropic has shipped 2 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 |
|---|---|---|---|
| Claude Fable 5.1xhigh | 1 Sept 2026 | 69.9 | #6 |
| Claude Fable 5thinkingthis page | 9 Jun 2026 | 69.2 | #13 |
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 | – | – | 53 tok/s | 3.75 s | $0.019 | Maths 67 · Reasoning 62 |
| medium | – | – | 53 tok/s | 3.75 s | $0.019 | Reasoning 68 |
| high | 64.9 | #46 | 53 tok/s | 3.75 s | $0.019 | Agents & tools 65 · Coding 70 · Human preference 71 · Maths 70 · Multimodal 68 · Reasoning 67 |
| xhigh | – | – | 53 tok/s | 3.75 s | $0.019 | Knowledge 74 · Reasoning 70 |
| max | 64.5 | #50 | 53 tok/s | 3.75 s | $0.019 | Agents & tools 51 · Coding 74 · Composite 73 · Maths 73 · Reasoning 70 |
| thinkingbest | 69.2 | #13 | 66 tok/s | 97 s | $0.019 | Agents & tools 74 · Composite 86 · Instruction following 63 · Knowledge 81 · Long context 67 · Reasoning 87 |
| not stated | 67.8 | #21 | 53 tok/s | 3.75 s | $0.019 | Agents & tools 67 · Coding 71 · Human preference 71 · Knowledge 67 · Reasoning 80 |
How its index has moved
8 Sept 2026 to 5 Oct 2026- thinking
- not stated
- high
- max
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
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 | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | 78.8%#126 | – | 83.3%#97 | – | 85.9%#78 | – | – | Epoch AI Benchmarking Hub | |
| SimpleBench | – | – | – | – | 81.9%#1 | – | – | SimpleBench | |
| LMArena Hard Prompts | – | – | 1530#4 | – | – | – | – | LMArena | |
| LiveBench Reasoningnot in index | – | – | – | – | 89.7%#14 | – | – | LiveBench | |
| GPQA Diamond (AA)not in index | – | – | – | – | – | 92.6%#32 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | – | – | – | – | 55.5%#8 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | – | – | 93.2%#10 | Vals AI | |
| EnigmaEval | – | – | 39.3%#1 | – | – | – | – | Scale AI SEAL | |
| Kagi LLM Benchmark | – | – | – | – | – | 91.4%#1 | 88.8%#2 | Kagi LLM Benchmark | |
| ARC-AGI-1 | 90.5%#65 | 92.5%#48 | 95.5%#29 | 98.5%#1 | 98.5%#1 | – | – | ARC Prize | |
| ARC-AGI-2 | 76.8%#42 | 82.5%#37 | 87.5%#22 | 88.3%#20 | 89.2%#17 | – | – | ARC Prize | |
| CritPt | – | – | – | – | – | 28.6%#23 | – | Artificial Analysis | |
| Chess Puzzlesnot in index | 28.0%#56 | – | 41.0%#22 | – | 41.0%#22 | – | – | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | – | – | – | – | 52.0%#10 | – | – | Epoch AI Benchmarking Hub | |
| LMCAnot in index | 59.6%#21 | 60.3%#20 | 61.1%#17 | 61.0%#18 | 60.3%#19 | – | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | 91.5%#57 | 95.7%#26 | 96.5%#17 | 97.3%#8 | 98.4%#2 | – | – | Epoch AI Benchmarking Hub | |
| BTF-3not in index | – | – | 13.0%#12 | – | – | – | – | Epoch AI Benchmarking Hub |
Coding
Agents & tools
Maths
| Benchmark | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | 87.0%#9 | – | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | 90.2%#7 | – | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 97.8%#24 | – | 100.0%#1 | – | 99.7%#12 | – | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | 95.0%#8 | – | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | – | – | 96.0%#11 | – | – | LiveBench | |
| LMArena Maths | – | – | 1522#2 | – | – | – | – | LMArena |
Knowledge
| Benchmark | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | 70.7%#6 | – | – | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | – | – | 80.5%#6 | – | – | LiveBench | |
| AA-Omniscience | – | – | – | – | – | 43.3#6 | – | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | – | – | 91.5%#3 | Vals AI | |
| LegalBench | – | – | – | – | – | – | 88.6%#1 | Vals AI | |
| CorpFin | – | – | – | – | – | – | 71.8%#2 | Vals AI | |
| TaxEval | – | – | – | – | – | – | 76.9%#5 | Vals AI | |
| PRBench Finance | – | – | – | – | – | – | 53.9%#3 | Scale AI SEAL | |
| PRBench Legal | – | – | – | – | – | – | 52.6%#3 | Scale AI SEAL | |
| Excel Modeling Benchmarknot in index | – | – | – | – | – | – | 73.7%#5 | Vals AI | |
| MedCodenot in index | – | – | – | – | – | – | 56.1%#4 | Vals AI | |
| MedScribenot in index | – | – | – | – | – | – | 88.5%#10 | Vals AI | |
| Public Benefits Benchnot in index | – | – | – | – | – | – | 70.4%#4 | Vals AI | |
| EBR-benchnot in index | – | – | – | – | 39.5%#8 | – | – | Epoch AI Benchmarking Hub | |
| GDP.pdfnot in index | – | – | – | – | 29.8%#6 | – | 30.0%#5 | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | – | – | – | – | – | 65.3%#5 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | – | – | – | – | – | 36.4%#167 | – | Artificial Analysis |
Instruction following
| Benchmark | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | – | – | 90.7%#1 | – | – | LiveBench | |
| LiveBench Instruction Followingnot in index | – | – | – | – | 75.8%#6 | – | – | LiveBench | |
| IFBench | – | – | – | – | – | 63.5%#112 | – | Artificial Analysis | |
| LMArena Instruction Followingnot in index | – | – | 1509#4 | – | – | – | – | LMArena |
Human preference
| Benchmark | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | 1504#3 | – | – | – | – | LMArena | |
| EQ-Bench 4 | – | – | – | – | – | – | 1340#2 | EQ-Bench | |
| LMArena Creative Writingnot in index | – | – | 1503#3 | – | – | – | – | LMArena | |
| LMArena Multi-turnnot in index | – | – | 1518#4 | – | – | – | – | LMArena | |
| LMArena Longer Queriesnot in index | – | – | 1522#4 | – | – | – | – | LMArena |
Multimodal
| Benchmark | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Vision | – | – | 1325#1 | – | – | – | – | LMArena | |
| MMMU-Pro (Vals)not in index | – | – | – | – | – | – | 89.3%#3 | Vals AI | |
| MortgageTaxnot in index | – | – | – | – | – | – | 68.9%#8 | Vals AI | |
| SAGEnot in index | – | – | – | – | – | – | 51.9%#9 | Vals AI | |
| Blueprint-Bench 2not in index | – | – | – | – | – | – | 38.6%#5 | Epoch AI Benchmarking Hub |
Long context
| Benchmark | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| AA-LCR | – | – | – | – | – | 82.3%#35 | – | Artificial Analysis | |
| LMArena Documentnot in index | – | – | – | – | – | – | 1499#1 | LMArena | |
| MLCRnot in index | – | – | – | – | – | 64.4%#4 | – | Artificial Analysis |
Composite
| Benchmark | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | – | 162.2#7 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | – | – | 83.0%#3 | – | – | LiveBench | |
| AA Intelligence Index | – | – | – | – | – | 49.6#19 | – | Artificial Analysis | |
| Vals Indexnot in index | – | – | – | – | – | – | 61.4#7 | Vals AI | |
| Vals Multimodal Indexnot in index | – | – | – | – | – | – | 74.2%#1 | Vals AI |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
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 | 59 tok/s | 4.56 s | $10.00 | $50.00 | 1M | – |
| Amazon Bedrock | 38 tok/s | 5.37 s | $10.00 | $50.00 | 1M | – |
| Google Vertex | 34 tok/s | 3.02 s | $10.00 | $50.00 | 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 $1.00 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.019 | $0.016 | 1.7 min |
| Summarise a 30-page report | 12,000 / 600 | $0.150 | $0.069 | 1.8 min |
| Code edit | 6,000 / 1,500 | $0.135 | $0.095 | 2.0 min |
| Agentic coding session | 60,000 / 4,000 | $0.800 | $0.395 | 2.6 min |
| Structured extraction | 2,000 / 200 | $0.030 | $0.017 | 1.7 min |
See also
Data as of 5 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Claude Fable 5: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/claude-fable-5, data as of 5 Oct 2026.