Claude Fable 5.1 vs Claude Opus 5.5
Full profiles: Claude Fable 5.1 (Anthropic) and Claude Opus 5.5 (Anthropic): every benchmark, price and rank.
Updated · Data from 10 sources
As of 21 Oct 2026, Claude Fable 5.1 (max) and Claude Opus 5.5 (max) are neck and neck on quality (70.6 vs 72.7 on the BenchLeader Index, inside the ±7.2 margin of error), and Claude Opus 5.5 (max) costs 2.5× less, which makes it the better buy for most workloads. Pick Claude Fable 5.1 for reasoning (77 vs 76); pick Claude Opus 5.5 for coding (82 vs 75).
Claude Fable 5.1 benchmarks max Anthropic Claude Opus 5.5 benchmarks max AnthropicOpen in the comparison toolWhich to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Claude Opus 5.5 | 82 vs 75 category score; Arena WebDev 1807 vs 1743 | Medium |
| Agents & tool use | Claude Opus 5.5 | 64 vs 61 category score; Terminal-Bench 64.8% vs 57.9% | Medium |
| Maths | Claude Opus 5.5 | 74 vs 73 category score; FrontierMath Tiers 1–3 91.2% vs 90.2% | Medium |
| Reasoning | Claude Fable 5.1 | 77 vs 76 category score; ARC-AGI-2 90.0% vs 91.7% | Medium |
| Knowledge | Claude Opus 5.5 | 79 vs 78 category score; SimpleQA Verified 72.2% vs 70.8% | Medium |
| Long documents | Either | 68 vs 68 category score; AA-LCR 85.3% vs 84.7% | Either |
| Images and other media | Claude Opus 5.5 | 71 vs 67 category score; AA-3D 1191 vs 1133 | Medium |
| Chat quality (human preference) | – | Only Claude Fable 5.1 has an Arena rating (70 category score) | Not comparable |
| Lowest cost | Claude Opus 5.5 | $8.00 vs $20.00 per 1M blended; agentic coding session $0.320 vs $0.800 | – |
| Fastest first answer | Claude Fable 5.1 | 221.4 s (Claude Fable 5.1 (max)) vs 608.7 s (Claude Opus 5.5 (max)) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Claude Fable 5.1 (max) and Claude Opus 5.5 (max) are neck and neck on quality (70.6 vs 72.7, within the ±7.2 margin).
- Claude Fable 5.1 (max) is stronger in human preference, long context, reasoning.
- Claude Opus 5.5 (max) is stronger in agents & tools, coding, composite, knowledge, maths, multimodal.
- Claude Opus 5.5 (max) is 2.5× cheaper ($8.00 vs $20.00 per 1M blended).
- Claude Opus 5.5 (max) streams 1.4× faster (95 vs 69 tokens per second).
At a glance
| Spec | Claude Fable 5.1 | Claude Opus 5.5 |
|---|---|---|
| Developer | Anthropic | Anthropic |
| Released | 1 Sept 2026 | 22 Sept 2026 |
| Context window (tokens) | 1,000,000 | 1,000,000 |
| Max output (tokens) | – | 128,000 (up to 300K on the Message Batches API (beta)) |
| Input types | – | text, image |
| Reasoning effort settings | – | low, medium, high, xhigh, max |
| Compared at | Claude Fable 5.1 (max) | Claude Opus 5.5 (max) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $10.00 / $50.00 / $0.250 | $4.00 / $20.00 / $0.200 |
| API model ID | – | claude-opus-5-5 |
| Where to use it | – | Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry |
Limits, availability and announced prices from the vendors' own pages: docs.anthropic.com (checked 20 Oct 2026). Everything else is from the day's data.
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Claude Fable 5.1 (max) | Claude Opus 5.5 (max) | Leader |
|---|---|---|---|
| Reasoning | 76.9 | 75.8 | Claude Fable 5.1 |
| Coding | 75.0 | 82.2 | Claude Opus 5.5 |
| Agents & tools | 60.8 | 63.5 | Claude Opus 5.5 |
| Maths | 72.9 | 74.4 | Claude Opus 5.5 |
| Knowledge | 77.5 | 78.7 | Claude Opus 5.5 |
| Human preference | 70.2 | – | not comparable |
| Multimodal | 67.1 | 71.0 | Claude Opus 5.5 |
| Long context | 67.9 | 67.5 | level |
| Composite | 89.4 | 94.2 | Claude Opus 5.5 |
Where sources disagree
Across the 48 benchmarks both have, Claude Fable 5.1 leads 9, Claude Opus 5.5 leads 34 and 5 are tied. Every overall index that rates both puts Claude Opus 5.5 ahead.
| Overall index | Claude Fable 5.1 (max) | Claude Opus 5.5 (max) | Leader |
|---|---|---|---|
| BenchLeader Index | 70.6 | 72.7 | neck and neck |
| AA Intelligence Index v4.3.2 | 53.4 | 57.6 | Claude Opus 5.5 |
| Arena Text | 1502 | 1508† | Claude Opus 5.5 |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.
| Measure | Claude Fable 5.1 (max) | Claude Opus 5.5 (max) |
|---|---|---|
| Blended price per 1M tokens | $20.00 | $8.00 |
| Input / output per 1M | $10.00 / $50.00 | $4.00 / $20.00 |
| Cached input per 1M | $0.250 | $0.200 |
| Price per task: chat reply | $0.019 | $0.0076 |
| Price per task: summarize a 30-page report | $0.150 | $0.060 |
| Price per task: code edit | $0.135 | $0.054 |
| Price per task: agentic coding session | $0.800 | $0.320 |
| Price per task: structured extraction | $0.030 | $0.012 |
| Cost to run a full benchmark suite | $7.63 | $5.98 |
| Output speed | 69 tok/s | 95 tok/s |
| Time to first answer | 221.4 s | 608.7 s |
| Time to a complete answer | 228.7 s | 613.9 s |
Per token is not per task: on Artificial Analysis' full suite Claude Fable 5.1 (max) cost $7.63 and Claude Opus 5.5 (max) $5.98, because Claude Opus 5.5 (max) used about 2.0× as many tokens for the same work, which a per-token price hides.
Each model is compared at its best-ranked reasoning-effort setting. Claude Fable 5.1 (max) scores 70.6 and takes 221 s to its first answer token; Claude Fable 5.1 (xhigh) scores 69.5 in 83 s, Claude Fable 5.1 (high) scores 69.4 in 21 s. Claude Opus 5.5 (max) scores 72.7 and takes 609 s to its first answer token; Claude Opus 5.5 (xhigh) scores 70.1 in 114 s, Claude Opus 5.5 (high) scores 69.4 in 34 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
127 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | Claude Fable 5.1 (max) | Claude Opus 5.5 (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 70.6 | 72.7 |
| Agents & tools score | BenchLeader | 60.8 | 63.5 |
| Coding score | BenchLeader | 75.0 | 82.2 |
| Composite score | BenchLeader | 89.4 | 94.2 |
| Human preference score | BenchLeader | 70.2 | – |
| Knowledge score | BenchLeader | 77.5 | 78.7 |
| Long context score | BenchLeader | 67.9 | 67.5 |
| Maths score | BenchLeader | 72.9 | 74.4 |
| Multimodal score | BenchLeader | 67.1 | 71.0 |
| Reasoning score | BenchLeader | 76.9 | 75.8 |
| Blended price $/M | BenchLeader | $20.00 | $8.00 |
| Output speed | BenchLeader | 69 tok/s | 95 tok/s |
| Time to first answer | BenchLeader | 221.4 s | 608.7 s |
| Context window | BenchLeader | 1,000,000 tokens | 1,000,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 90.6% |
| Humanity's Last Exam | Scale AI / CAIS | 46.5%† | – |
| Arena Hard Prompts | Arena (LMArena) | 1521 | 1533† |
| LiveBench 2.0 Reasoning | LiveBench | 91.7%† | 92.2%† |
| LiveBench 1.0 Reasoning | LiveBench | 91.7% | 92.2% |
| GPQA Diamond (AA) | Artificial Analysis | 93.7% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 59.1% | 61.4% |
| GPQA Diamond (Vals) | Vals AI | 93.4%† | – |
| ARC-AGI-1 | ARC Prize | 97.5% | 97.5% |
| ARC-AGI-2 | ARC Prize | 90.0% | 91.7% |
| CritPt | Artificial Analysis | 29.7% | 31.7% |
| MysteryMechanism | Vals AI | 47.8%† | 49.5%† |
| Chess Puzzles | Epoch AI Benchmarking Hub | 48.0% | 67.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 58.0% | 71.0% |
| LMCA | Epoch AI Benchmarking Hub | 65.5%† | 68.2% |
| DTBench | Epoch AI Benchmarking Hub | 97.6%† | 98.9% |
| Coding | |||
| SciCode | SciCode | 63.1% | 66.9% |
| WeirdML | WeirdML | 92.9% | – |
| FrontierCode | Cognition | 50.9%† | 54.6%† |
| GSO-Bench | GSO-Bench | 88.2%† | – |
| Arena Coding | Arena (LMArena) | 1537 | 1548† |
| Arena WebDev | Arena (LMArena) | 1743 | 1807 |
| LiveBench 2.0 Coding | LiveBench | 97.7%† | 96.3%† |
| LiveBench 1.0 Coding | LiveBench | 86.4% | 89.3% |
| LiveCodeBench | Vals AI | 90.5%† | – |
Show all 127 benchmarks
| Metric | Publisher | Claude Fable 5.1 (max) | Claude Opus 5.5 (max) |
|---|---|---|---|
| Coding | |||
| IOI | Vals AI | 90.8%† | 95.1%† |
| Code Migration | Vals AI | 54.6%† | 66.7%† |
| ProgramBench | Vals AI | 7.0%† | 18.5%† |
| Vibe Code Bench 1-100 | Vals AI | 28.0%† | 30.4%† |
| Vibe Code Bench v1.1 | Vals AI | 90.3%† | 90.3%† |
| SWE Atlas: Codebase QnA | Scale AI SEAL | 60.0%† | – |
| SWE Atlas: Refactoring | Scale AI SEAL | 56.7%† | – |
| SWE Atlas: Test Writing | Scale AI SEAL | 67.0%† | – |
| MirrorCode | Epoch AI Benchmarking Hub | 73.3%† | 77.4% |
| CursorBench | Cursor | 51.8% | 57.8% |
| ALE-Bench | Epoch AI Benchmarking Hub | 2143.2† | 2147.4† |
| FrontierSWE | Epoch AI Benchmarking Hub | 56.3% | 62.3% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 52.0% | 59.6% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 91.4% | – |
| Startrise Frontend | Startrise | 77.1%† | 71.1%† |
| FrontierCode 1.1 (AA) | Artificial Analysis | 49.0% | 52.2% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 57.9% | 64.8% |
| Remote Labor Index | Scale AI / CAIS | 17.9%† | 21.3%† |
| APEX-Agents | Mercor | 68.6%† | 73.5% |
| Arena Agent | Arena (LMArena) | 14.3 | 20.4 |
| LiveBench Long Agentic | LiveBench | 80.8%† | 85.0%† |
| LiveBench Browser | LiveBench | 100.0%† | 80.0%† |
| LiveBench Agentic Coding | LiveBench | 66.1% | 71.7% |
| Terminal-Bench 2.1 (Vals) | Vals AI | 85.0%† | 87.6%† |
| MCP Atlas | Scale AI SEAL | 87.2%† | – |
| HiL-Bench | Scale AI SEAL | 61.5%† | 59.0%† |
| GDPval-AA v2.1 | Artificial Analysis | 62.7% | 68.3% |
| τ³-Banking (AA) | Artificial Analysis | 47.2% | – |
| ITBench SRE (AA) | Artificial Analysis | 49.5% | 38.2% |
| Analyst Agent (AA) | Artificial Analysis | 57.5% | 56.3% |
| CUA-bench | Vals AI | 13.2% | 14.0% |
| CyberBench | Vals AI | 70.4%† | 55.4%† |
| Finance Agent v2 | Vals AI | 58.9%† | 58.6%† |
| Harvey's Legal Agent Benchmark | Vals AI | – | 9.2%† |
| Legal Research Bench | Vals AI | 55.3%† | 50.5%† |
| SkillsBench | Vals AI | 61.5%† | – |
| SREBench | Vals AI | 22.9%† | 33.6%† |
| Tax Agent Bench | Vals AI | 49.2%† | 45.6%† |
| Terminal-Bench 4.0 (Vals) | Vals AI | 58.1%† | 65.2%† |
| Terminal-Bench Science | Vals AI | 40.0%† | 47.1%† |
| Time Horizon Index: KSP | Vals AI | 63.3%† | 91.3%† |
| Arena Search | Arena (LMArena) | 1233† | 1211† |
| Vending-Bench 2 | Andon Labs | 5421.6† | 9235.3† |
| AutomationBench | Artificial Analysis | 59.4% | 69.5% |
| GDP.pdf (AA) | Artificial Analysis | 26.2% | 26.2% |
| Harvey LAB-AA v1.1 | Harvey | 6.4% | 4.2% |
| AA-Briefcase v1.1 | Artificial Analysis | 1672 | 1803 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 43.3% | 59.0% |
| CWE-Bench (AA) | Artificial Analysis | 50.8% | 58.3% |
| DeepSec-Bench (AA) | Artificial Analysis | 23.5% | 26.9% |
| CyberGym end-to-end (AA) | Artificial Analysis | 0.8% | 0.8% |
| FACTS Search | – | 86.5%† | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 90.2% | 91.2% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 87.8% | 95.0% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 100.0% | 100.0% |
| ProofBench | Vals AI | 100.0% | 100.0% |
| LiveBench Mathematics | LiveBench | 97.0% | 97.1% |
| Arena Maths | Arena (LMArena) | 1525 | 1502† |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 70.8% | 72.2% |
| LiveBench 2.0 Data Analysis | LiveBench | 80.3%† | 80.3%† |
| LiveBench Research | LiveBench | 82.0%† | 85.0%† |
| LiveBench 1.0 Data Analysis | LiveBench | 80.3% | 80.3% |
| AA-Omniscience | Artificial Analysis | 43.5 | 46.4 |
| MMLU-Pro | Vals AI | 92.4%† | – |
| LegalBench | Vals AI | 88.5%† | – |
| TaxEval | Vals AI | 76.0%† | – |
| PRBench Finance | Scale AI SEAL | 50.8%† | – |
| PRBench Legal | Scale AI SEAL | 51.6%† | 47.6%† |
| BioMysteryBench | Vals AI | – | 79.3%† |
| Excel Modeling Benchmark | Vals AI | 76.7%† | 75.9%† |
| MedCode | Vals AI | 53.5%† | 49.8%† |
| MedScribe | Vals AI | 91.3%† | 91.4%† |
| Public Benefits Bench | Vals AI | 74.9%† | 70.6%† |
| EBR-bench | Epoch AI Benchmarking Hub | 57.1% | 71.4% |
| GDP.pdf | Epoch AI Benchmarking Hub | 27.6% | 30.6% |
| AA-Omniscience: accuracy | Artificial Analysis | 67.2% | 66.2% |
| AA-Omniscience: non-hallucination | Artificial Analysis | 27.4% | 41.4% |
| Instruction following | |||
| LiveBench Language | LiveBench | 89.5% | 86.3% |
| LiveBench Instruction Following | LiveBench | 73.0% | 65.7% |
| Arena Instruction Following | Arena (LMArena) | 1499 | 1518† |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1502 | 1508† |
| Arena Creative Writing | Arena (LMArena) | 1485 | 1519† |
| Arena Multi-turn | Arena (LMArena) | 1487 | 1501† |
| Arena Longer Queries | Arena (LMArena) | 1509 | 1523† |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1320 | 1321† |
| LiveBench Design | LiveBench | 83.4%† | 86.9%† |
| MMMU-Pro | Artificial Analysis | – | 87.7% |
| MMMU-Pro (Vals) | Vals AI | 90.6%† | – |
| MortgageTax | Vals AI | 70.8%† | – |
| SAGE | Vals AI | 48.5%† | 45.8%† |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 41.9%† | 51.2%† |
| AA-3D | Artificial Analysis | 1133 | 1191 |
| Long context | |||
| AA-LCR | Artificial Analysis | 85.3% | 84.7% |
| Arena Document | Arena (LMArena) | 1493 | – |
| MLCR | Artificial Analysis | 71.1% | 66.7% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 166.5† | 168.8† |
| LiveBench 2.0 | LiveBench | 88.0%† | 86.5%† |
| LiveBench 1.0 | LiveBench | 83.4% | 83.2% |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 53.4 | 57.6 |
| Vals Index | Vals AI | – | 70.2† |
| AA Cyber Index | Artificial Analysis | 25.0 | 28.7 |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 13.7%† | 5.8% |
Data as of 2026-10-21. Best configuration of each model; every score links to its source on the model pages.
What changed
- Claude Fable 5.1 (max): 1 score revision.
- Claude Opus 5.5 (max): 2 score revisions.
- Claude Fable 5.1 (max): BenchLeader Index 70.0 → 70.6.
- Claude Opus 5.5 (max): BenchLeader Index 71.9 → 72.6.
- Claude Opus 5.5 (max): 1 new result (FORTRESS).
- Claude Opus 5.5 (max): BenchLeader Index 71.3 → 71.9.
- Claude Opus 5.5 (max): 1 score revision.
- Claude Fable 5.1 (max): 2 score revisions.
- Claude Opus 5.5 (max): 1 new result (Chess Puzzles); 1 score revision.
- Claude Fable 5.1 (max): 1 score revision.
Related comparisons
- Claude Fable 5.1 vs Gemini 4 Argon
- Claude Fable 5.1 vs Muse Wonder Full
- Claude Fable 5.1 vs GPT-6 Astra
- Claude Opus 5.5 vs Gemini 4 Argon
- Claude Opus 5.5 vs Muse Wonder Full
- Claude Opus 5.5 vs GPT-6 Astra
- Claude Opus 5 vs Claude Opus 5.5
- Claude Fable 5.1 vs Claude Opus 5
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
Claude Fable 5.1 vs Claude Opus 5.5: questions
- Is Claude Fable 5.1 better than Claude Opus 5.5?
- Claude Fable 5.1 (max) and Claude Opus 5.5 (max) are neck and neck on quality (70.6 vs 72.7, within the ±7.2 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2026-10-21, so check the category scores for your use.
- Which is cheaper, Claude Fable 5.1 or Claude Opus 5.5, and will that change?
- Claude Opus 5.5 is cheaper today: $8.00 against $20.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Claude Fable 5.1 better than Claude Opus 5.5 for coding?
- Claude Opus 5.5 scores higher in coding (82 vs 75 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1743 for Claude Fable 5.1 against 1807 for Claude Opus 5.5.
- Is Claude Fable 5.1 better than Claude Opus 5.5 for agents and tool use?
- Claude Opus 5.5 scores higher in agentic tasks (64 vs 61 on the category index, where 50 is average). On Terminal-Bench, the top shared agents & tools benchmark, it is 57.9% for Claude Fable 5.1 against 64.8% for Claude Opus 5.5.
- Which is better for long documents?
- They are level on long-context benchmarks (68 vs 68). Both accept 1,000,000 tokens of context.
- Which has lower latency?
- Claude Fable 5.1 (max) answers first: 221.4 s against 608.7 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Which is faster, Claude Fable 5.1 or Claude Opus 5.5?
- Claude Opus 5.5 streams faster: 95 against 69 output tokens per second.
- Do independent benchmarks agree on Claude Fable 5.1 vs Claude Opus 5.5?
- Across the 48 benchmarks both have, Claude Fable 5.1 leads 9, Claude Opus 5.5 leads 34 and 5 are tied. Every overall index that rates both puts Claude Opus 5.5 ahead.
- When was Claude Opus 5.5 released?
- Claude Opus 5.5 was released on 22 Sept 2026.