Claude Fable 5.1 vs GPT-5.2
Full profiles: Claude Fable 5.1 (Anthropic) and GPT-5.2 (OpenAI): every benchmark, price and rank.
Updated · Data from 10 sources
As of 26 Oct 2026, Claude Fable 5.1 (max) leads on quality (70.5 vs 60.7 on the BenchLeader Index), but GPT-5.2 (xhigh) costs 4.2× less, so it is the better buy unless you need the extra quality. Claude Fable 5.1 leads in every category we measure both in.
Claude Fable 5.1 benchmarks max Anthropic
GPT-5.2 benchmarks xhigh OpenAIOpen 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 Fable 5.1 | 75 vs 63 category score; WeirdML 92.9% vs 72.2% | High |
| Agents & tool use | Claude Fable 5.1 | 61 vs 55 category score | Medium |
| Maths | Claude Fable 5.1 | 73 vs 57 category score; FrontierMath Tiers 1–3 90.2% vs 67.4% | High |
| Reasoning | Claude Fable 5.1 | 77 vs 61 category score; ARC-AGI-2 90.0% vs 52.9% | High |
| Knowledge | Claude Fable 5.1 | 78 vs 57 category score; SimpleQA Verified 70.8% vs 37.1% | High |
| Long documents | Claude Fable 5.1 | 68 vs 66 category score; AA-LCR 85.3% vs 82.7% | Medium |
| Chat quality (human preference) | – | Only Claude Fable 5.1 has an Arena rating (70 category score) | Not comparable |
| Lowest cost | GPT-5.2 | $4.81 vs $20.00 per 1M blended; agentic coding session $0.161 vs $0.800 | – |
| Fastest first answer | GPT-5.2 | 124.7 s (GPT-5.2 (xhigh)) vs 180.5 s (Claude Fable 5.1 (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) leads on quality: 70.5 vs 60.7.
- Claude Fable 5.1 (max) is stronger in agents & tools, coding, composite, human preference, knowledge, long context, maths, multimodal, reasoning.
- GPT-5.2 (xhigh) is stronger in instruction following.
- GPT-5.2 (xhigh) is 4.2× cheaper ($4.81 vs $20.00 per 1M blended).
At a glance
| Spec | Claude Fable 5.1 | GPT-5.2 |
|---|---|---|
| Developer | Anthropic | OpenAI |
| Released | 1 Sept 2026 | 11 Dec 2025 |
| Context window (tokens) | 1,000,000 | 400,000 |
| Compared at | Claude Fable 5.1 (max) | GPT-5.2 (xhigh) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $10.00 / $50.00 / $0.250 | $1.75 / $14.00 / $0.175 |
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) | GPT-5.2 (xhigh) | Leader |
|---|---|---|---|
| Reasoning | 76.8 | 60.7 | Claude Fable 5.1 |
| Coding | 74.9 | 62.9 | Claude Fable 5.1 |
| Agents & tools | 60.5 | 54.9 | Claude Fable 5.1 |
| Maths | 72.9 | 56.8 | Claude Fable 5.1 |
| Knowledge | 77.5 | 57.1 | Claude Fable 5.1 |
| Instruction following | – | 73.6 | not comparable |
| Human preference | 70.0 | – | not comparable |
| Multimodal | 67.1 | – | not comparable |
| Long context | 67.8 | 66.4 | Claude Fable 5.1 |
| Composite | 89.2 | 63.5 | Claude Fable 5.1 |
Where sources disagree
Across the 18 benchmarks both have, Claude Fable 5.1 leads 17, GPT-5.2 leads 0 and 1 is a tie. Every overall index that rates both puts Claude Fable 5.1 ahead.
| Overall index | Claude Fable 5.1 (max) | GPT-5.2 (xhigh) | Leader |
|---|---|---|---|
| BenchLeader Index | 70.5 | 60.7 | Claude Fable 5.1 |
| AA Intelligence Index v4.3.2 | 53.4 | 30.4 | Claude Fable 5.1 |
| Arena Text | 1503 | 1437† | Claude Fable 5.1 |
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) | GPT-5.2 (xhigh) |
|---|---|---|
| Blended price per 1M tokens | $20.00 | $4.81 |
| Input / output per 1M | $10.00 / $50.00 | $1.75 / $14.00 |
| Cached input per 1M | $0.250 | $0.175 |
| Price per task: chat reply | $0.019 | $0.0049 |
| Price per task: summarize a 30-page report | $0.150 | $0.029 |
| Price per task: code edit | $0.135 | $0.032 |
| Price per task: agentic coding session | $0.800 | $0.161 |
| Price per task: structured extraction | $0.030 | $0.0063 |
| Cost to run a full benchmark suite | $7.63 | – |
| Output speed | 70 tok/s | 81 tok/s |
| Time to first answer | 180.5 s | 124.7 s |
| Time to a complete answer | 187.7 s | 130.9 s |
Each model is compared at its best-ranked reasoning-effort setting. Claude Fable 5.1 (max) scores 70.5 and takes 181 s to its first answer token; Claude Fable 5.1 (xhigh) scores 69.5 in 91 s, Claude Fable 5.1 (high) scores 69.4 in 18 s. GPT-5.2 (xhigh) scores 60.7 and takes 125 s to its first answer token; GPT-5.2 scores 58.0 in 2 s, GPT-5.2 (medium) scores 57.6 in 2 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
167 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) | GPT-5.2 (xhigh) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 70.5 | 60.7 |
| Agents & tools score | BenchLeader | 60.5 | 54.9 |
| Coding score | BenchLeader | 74.9 | 62.9 |
| Composite score | BenchLeader | 89.2 | 63.5 |
| Human preference score | BenchLeader | 70.0 | – |
| Knowledge score | BenchLeader | 77.5 | 57.1 |
| Long context score | BenchLeader | 67.8 | 66.4 |
| Maths score | BenchLeader | 72.9 | 56.8 |
| Multimodal score | BenchLeader | 67.1 | – |
| Reasoning score | BenchLeader | 76.8 | 60.7 |
| Instruction following score | BenchLeader | – | 73.6 |
| Blended price $/M | BenchLeader | $20.00 | $4.81 |
| Output speed | BenchLeader | 70 tok/s | 81 tok/s |
| Time to first answer | BenchLeader | 180.5 s | 124.7 s |
| Context window | BenchLeader | 1,000,000 tokens | 400,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 91.4% |
| Humanity's Last Exam | Scale AI / CAIS | 46.5%† | 27.8%† |
| SimpleBench | SimpleBench | – | 45.8%† |
| Arena Hard Prompts | Arena (LMArena) | 1523 | 1497† |
| LiveBench 2.0 Reasoning | LiveBench | 91.7%† | – |
| LiveBench 1.0 Reasoning | LiveBench | 91.7% | 83.2%† |
| GPQA Diamond (AA) | Artificial Analysis | 93.7% | 90.3% |
| Humanity's Last Exam (AA) | Artificial Analysis | 59.1% | 37.7% |
| GPQA Diamond (Vals) | Vals AI | 93.4%† | 91.7% |
| Kagi LLM Benchmark | Kagi LLM Benchmark | – | 73.3%† |
| ARC-AGI-1 | ARC Prize | 97.5% | 86.2% |
| ARC-AGI-2 | ARC Prize | 90.0% | 52.9% |
| CritPt | Artificial Analysis | 29.7% | 11.6% |
| MysteryMechanism | Vals AI | 47.8%† | – |
| SciPredict | Scale AI SEAL | – | 20.6%† |
| Chess Puzzles | Epoch AI Benchmarking Hub | 48.0% | 48.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 58.0% | 22.0%† |
| CL-bench | Epoch AI Benchmarking Hub | – | 18.2%† |
| LMCA | Epoch AI Benchmarking Hub | 65.5%† | 43.9% |
| DTBench | Epoch AI Benchmarking Hub | 97.6%† | 90.9% |
| ForecastBench | ForecastBench | – | 64.2%† |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | – | 73.8%† |
| SciCode | SciCode | 63.1% | – |
| WeirdML | WeirdML | 92.9% | 72.2% |
| FrontierCode | Cognition | 50.9%† | – |
Show all 167 benchmarks
| Metric | Publisher | Claude Fable 5.1 (max) | GPT-5.2 (xhigh) |
|---|---|---|---|
| Coding | |||
| GSO-Bench | GSO-Bench | 88.2%† | 27.4%† |
| Arena Coding | Arena (LMArena) | 1538 | 1515† |
| Arena WebDev | Arena (LMArena) | 1743 | 1417† |
| LiveBench 2.0 Coding | LiveBench | 97.7%† | – |
| LiveBench 1.0 Coding | LiveBench | 86.4% | 76.1%† |
| LiveCodeBench | Vals AI | 90.5%† | 85.4% |
| IOI | Vals AI | 90.8%† | – |
| SWE-bench (Vals) | Vals AI | – | 75.8% |
| SWE-Bench Pro | Scale AI SEAL | – | 29.9%† |
| SWE-bench Verified (bash only) | SWE-bench | – | 69.0%† |
| SWE-bench Verified (any scaffold) | SWE-bench | – | 72.8%† |
| Code Migration | Vals AI | 54.6%† | – |
| ProgramBench | Vals AI | 7.0%† | – |
| SWE-bench Multimodal | Vals AI | 61.3%† | – |
| Vibe Code Bench 1-100 | Vals AI | 28.0%† | – |
| Vibe Code Bench v1.1 | Vals AI | 90.3%† | 53.5% |
| SWE-Bench Pro (private) | Scale AI SEAL | – | 29.9%† |
| 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%† | – |
| CursorBench | Cursor | 51.8% | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 2143.2† | 1249.8† |
| AlgoTune | AlgoTune | – | 2.0† |
| FrontierSWE | Epoch AI Benchmarking Hub | 56.3% | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 52.0% | – |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 91.4% | – |
| Startrise Frontend | Startrise | 77.1%† | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | 49.0% | – |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 57.9% | 64.9%† |
| GDPval | OpenAI | – | 49.7%† |
| Remote Labor Index | Scale AI / CAIS | 17.9%† | 2.1%† |
| APEX-Agents | Mercor | 68.6%† | – |
| Arena Agent | Arena (LMArena) | 13.9 | – |
| LiveBench Long Agentic | LiveBench | 77.9%† | – |
| LiveBench Browser | LiveBench | 89.8%† | – |
| LiveBench Agentic Coding | LiveBench | 66.1% | 50.3%† |
| Terminal-Bench Hard | Artificial Analysis | – | 47.0% |
| τ²-Bench Telecom (AA) | Artificial Analysis | – | 84.8% |
| Terminal-Bench 2.1 (Vals) | Vals AI | 85.0%† | – |
| MCP Atlas | Scale AI SEAL | 87.2%† | 67.6% |
| HiL-Bench | Scale AI SEAL | 61.5%† | – |
| BFCL Overall | Berkeley Function Calling Leaderboard | – | 55.9%† |
| τ²-bench | τ²-bench | – | 84.8%† |
| τ³-bench Banking | τ²-bench | – | 32.2%† |
| GDPval-AA v2.1 | Artificial Analysis | 62.7% | – |
| τ³-Banking (AA) | Artificial Analysis | 47.2% | 11.1%† |
| ITBench SRE (AA) | Artificial Analysis | 49.5% | – |
| Analyst Agent (AA) | Artificial Analysis | 57.5% | – |
| CUA-bench | Vals AI | 13.2% | – |
| CyberBench | Vals AI | 70.4%† | – |
| Finance Agent v2 | Vals AI | 58.9%† | – |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 6.7%† | – |
| Legal Research Bench | Vals AI | 55.3%† | – |
| Poker Agent | Vals AI | – | 1131.8%† |
| Robo-bench | Vals AI | 5.8% | – |
| SkillsBench | Vals AI | 61.5%† | – |
| SREBench | Vals AI | 22.9%† | – |
| Tax Agent Bench | Vals AI | 49.2%† | – |
| Terminal-Bench 2.0 (Vals) | Vals AI | – | 51.7%† |
| Terminal-Bench 4.0 (Vals) | Vals AI | 58.1%† | – |
| Terminal-Bench Science | Vals AI | 40.0%† | – |
| Time Horizon Index: KSP | Vals AI | 63.3%† | – |
| Arena Search | Arena (LMArena) | 1233† | – |
| DeepResearch Bench | DeepResearch Bench | – | 41.1%† |
| METR Time Horizons | METR | – | 75.3%† |
| Vending-Bench 2 | Andon Labs | 5421.6† | 3591.3† |
| AutomationBench | Artificial Analysis | 59.4% | – |
| GDP.pdf (AA) | Artificial Analysis | 26.2% | – |
| Harvey LAB-AA v1.1 | Harvey | 6.4% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1671 | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 43.3% | – |
| CWE-Bench (AA) | Artificial Analysis | 50.8% | – |
| DeepSec-Bench (AA) | Artificial Analysis | 23.5% | – |
| CyberGym end-to-end (AA) | Artificial Analysis | 0.8% | – |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 90.2% | 67.4% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 87.8% | 31.7% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 100.0% | 96.1% |
| ProofBench | Vals AI | 100.0% | 15.0% |
| LiveBench Mathematics | LiveBench | 97.0% | 93.2%† |
| AIME (Vals) | Vals AI | – | 96.9% |
| MGSM | Vals AI | – | 94.0% |
| AIME 2026 | MathArena | – | 98.3%† |
| HMMT February 2026 | MathArena | – | 97.0%† |
| MathArena Apex | MathArena | – | 13.5%† |
| Arena Maths | Arena (LMArena) | 1529 | 1454† |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 70.8% | 37.1% |
| LiveBench 2.0 Data Analysis | LiveBench | 80.3%† | – |
| LiveBench Research | LiveBench | 82.0%† | – |
| LiveBench 1.0 Data Analysis | LiveBench | 80.3% | 78.2%† |
| AA-Omniscience | Artificial Analysis | 43.5 | -0.9 |
| MMLU-Pro | Vals AI | 92.4%† | 86.2% |
| LegalBench | Vals AI | 88.5%† | 82.8% |
| CorpFin | Vals AI | – | 65.9% |
| TaxEval | Vals AI | 76.0%† | 75.8% |
| MedQA | Vals AI | – | 94.1% |
| PRBench Finance | Scale AI SEAL | 50.8%† | – |
| PRBench Legal | Scale AI SEAL | 51.6%† | – |
| MultiNRC | Scale AI SEAL | – | 42.2%† |
| CaseLaw v2 | Vals AI | – | 66.0% |
| Excel Modeling Benchmark | Vals AI | 76.7%† | – |
| Excel Modeling Benchmark v1.1 | Vals AI | 76.7%† | – |
| MedCode | Vals AI | 53.5%† | 49.8% |
| MedScribe | Vals AI | 91.3%† | 84.4% |
| Public Benefits Bench v1.1 | Vals AI | 74.9%† | – |
| DrugDiscoveryBench | Scale AI SEAL | – | 29.3% |
| EBR-bench | Epoch AI Benchmarking Hub | 57.1% | 23.0% |
| GDP.pdf | Epoch AI Benchmarking Hub | 27.6% | – |
| AA-Omniscience: accuracy | Artificial Analysis | 67.2% | 44.3% |
| AA-Omniscience: non-hallucination | Artificial Analysis | 27.4% | 18.8% |
| Global-MMLU-Lite (AA) | Artificial Analysis | – | 89.8%† |
| Instruction following | |||
| LiveBench Language | LiveBench | 89.5% | 79.8%† |
| LiveBench Instruction Following | LiveBench | 73.0% | 61.8%† |
| IFBench | Artificial Analysis | – | 75.4% |
| TutorBench | Scale AI SEAL | – | 53.5%† |
| Arena Instruction Following | Arena (LMArena) | 1501 | 1457† |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1503 | 1476† |
| Arena Creative Writing | Arena (LMArena) | 1487 | 1433† |
| Arena Multi-turn | Arena (LMArena) | 1488 | 1495† |
| Arena Longer Queries | Arena (LMArena) | 1511 | 1472† |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1320 | 1268† |
| LiveBench Design | LiveBench | 83.4%† | – |
| MMMU-Pro | Artificial Analysis | – | 74.6%† |
| VISTA | Scale AI SEAL | – | 46.6%† |
| MMMU-Pro (Vals) | Vals AI | 90.6%† | 86.7% |
| MortgageTax | Vals AI | 70.8%† | 67.1% |
| SAGE | Vals AI | 48.5%† | 49.3% |
| VPCT | Epoch AI Benchmarking Hub | – | 84.0% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 41.9%† | – |
| AA-3D | Artificial Analysis | 1135 | – |
| Long context | |||
| AA-LCR | Artificial Analysis | 85.3% | 82.7% |
| Arena Document | Arena (LMArena) | 1493 | 1425† |
| MLCR | Artificial Analysis | 15.6% | – |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 166.5† | 153.7† |
| LiveBench 2.0 | LiveBench | 85.7%† | – |
| LiveBench 1.0 | LiveBench | 83.4% | 74.6%† |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 53.4 | 30.4 |
| Vals RSI Index | Vals AI | 36.1% | 15.8% |
| AA Cyber Index | Artificial Analysis | 25.0 | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 13.7%† | 17.5%† |
| MASK | Scale AI SEAL | – | 86.7%† |
| PropensityBench | Scale AI SEAL | – | 34.4%† |
Data as of 2026-10-26. Best configuration of each model; every score links to its source on the model pages.
What changed
- Claude Fable 5.1 (max): 2 new results (Robo-bench, Vals RSI Index).
- GPT-5.2 (xhigh): 1 new result (Vals RSI Index).
- Claude Fable 5.1 (max): 1 score revision.
- Claude Fable 5.1 (max): 1 score revision.
- Claude Fable 5.1 (max): BenchLeader Index 70.0 → 70.6.
- Claude Fable 5.1 (max): 2 score revisions.
- GPT-5.2 (xhigh): 1 score revision.
- Claude Fable 5.1 (max): 1 score revision.
- Claude Fable 5.1 (max): BenchLeader Index 70.9 → 70.0.
- Claude Fable 5.1 (max): 7 new results (FrontierCode 1.1 (AA), Terminal-Bench-Science 0.1 (AA), AA Cyber Index…).
Related comparisons
- Claude Fable 5.1 vs Claude Opus 5.5
- Claude Fable 5.1 vs Gemini 4 Argon
- Claude Fable 5.1 vs Muse Wonder Full
- GPT-5.2 vs Claude Opus 5.5
- GPT-5.2 vs Gemini 4 Argon
- GPT-5.2 vs Muse Wonder Full
- Claude Opus 5 vs GPT-5.2
- Claude Fable 5.1 vs GPT-5 Pro
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
Claude Fable 5.1 vs GPT-5.2: questions
- Is Claude Fable 5.1 better than GPT-5.2?
- Claude Fable 5.1 (max) leads on quality: 70.5 vs 60.7. The BenchLeader Index combines every independent quality benchmark; Claude Fable 5.1 (max) is ahead overall as of 2026-10-26, but check the category scores for your use.
- Which is cheaper, Claude Fable 5.1 or GPT-5.2, and will that change?
- GPT-5.2 is cheaper today: $4.81 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 GPT-5.2 for coding?
- Claude Fable 5.1 scores higher in coding (75 vs 63 on the category index, where 50 is average). On WeirdML, the top shared coding benchmark, it is 92.9% for Claude Fable 5.1 against 72.2% for GPT-5.2.
- Is Claude Fable 5.1 better than GPT-5.2 for agents and tool use?
- Claude Fable 5.1 scores higher in agentic tasks (61 vs 55 on the category index, where 50 is average).
- Which is better for long documents?
- Claude Fable 5.1 scores higher on long-context benchmarks (68 vs 66). Claude Fable 5.1 accepts more context: 1,000,000 against 400,000 tokens for GPT-5.2.
- Which has lower latency?
- GPT-5.2 (xhigh) answers first: 124.7 s against 180.5 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 GPT-5.2?
- GPT-5.2 streams faster: 81 against 70 output tokens per second.
- Do independent benchmarks agree on Claude Fable 5.1 vs GPT-5.2?
- Across the 18 benchmarks both have, Claude Fable 5.1 leads 17, GPT-5.2 leads 0 and 1 is a tie. Every overall index that rates both puts Claude Fable 5.1 ahead.
- When was Claude Fable 5.1 released?
- Claude Fable 5.1 was released on 1 Sept 2026.