Claude Fable 5.2 vs DeepSeek V4 Pro
Full profiles: Claude Fable 5.2 (Anthropic) and DeepSeek V4 Pro (DeepSeek): every benchmark, price and rank.
Updated · Data from 7 sources
As of 9 Nov 2026, Claude Fable 5.2 is ahead of DeepSeek V4 Pro on quality (69.4 vs 59.1 on the BenchLeader Index), but DeepSeek V4 Pro is 16× cheaper at $0.990 vs $16.00 per 1M tokens blended. Claude Fable 5.2 leads in every category we measure both in. Claude Fable 5.2 streams ~1.5× faster.
Key facts (as of 9 Nov 2026)
- Claude Fable 5.2 scores 69.4 on the BenchLeader Index and DeepSeek V4 Pro scores 59.1 (BenchLeader, 9 Nov 2026).
- Claude Fable 5.2 costs $16.00 per 1M tokens blended (list price, 9 Nov 2026).
- DeepSeek V4 Pro costs $0.990 per 1M tokens blended (list price, 9 Nov 2026).
- Claude Fable 5.2 streams ~1.5× faster than DeepSeek V4 Pro (Artificial Analysis, 7-day median, 9 Nov 2026).
- Of the 2 benchmarks both models have, Claude Fable 5.2 leads 2 (BenchLeader, 9 Nov 2026).
Claude Fable 5.2 benchmarks 
DeepSeek V4 Pro benchmarks Which 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 5.2 | 80 vs 54 category score | High |
| Reasoning | Claude Fable 5.2 | 78 vs 62 category score; ARC-AGI-2 92.5% vs 59.7% | High |
| Chat quality (human preference) | – | Only DeepSeek V4 Pro has an Arena rating (65 category score) | Not comparable |
| Lowest cost | DeepSeek V4 Pro | $0.990 vs $16.00 per 1M blended; agentic coding session $0.048 vs $0.640 | – |
| Fastest first answer | DeepSeek V4 Pro | 1.5 s (DeepSeek V4 Pro (high)) vs 3.0 s (Claude Fable 5.2 (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 5.2 (max) leads on quality: 69.4 vs 59.1.
- Claude Fable 5.2 (max) is stronger in coding, knowledge, long context, reasoning.
- DeepSeek V4 Pro (high) is stronger in human preference, maths.
- DeepSeek V4 Pro (high) is 16× cheaper ($0.990 vs $16.00 per 1M blended).
- Claude Fable 5.2 (max) streams 1.6× faster (72 vs 44 tokens per second).
At a glance
| Spec | Claude Fable 5.2 | DeepSeek V4 Pro |
|---|---|---|
| Developer | Anthropic | DeepSeek |
| Released | 9 Nov 2026 | 13 Aug 2026 |
| Context window (tokens) | 1,000,000 | 1,048,576 |
| Compared at | Claude Fable 5.2 (max) | DeepSeek V4 Pro (high) |
| Open weights | No | Yes |
| Price in / out / cached input per 1M | $8.00 / $40.00 / $0.200 | $0.660 / $1.98 / $0.022 |
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 5.2 (max) | DeepSeek V4 Pro (high) | Leader |
|---|---|---|---|
| Reasoning | 78.4 | 62.2 | Claude Fable 5.2 |
| Coding | 80.2 | 53.8 | Claude Fable 5.2 |
| Maths | – | 64.6 | not comparable |
| Knowledge | 81.7 | – | not comparable |
| Human preference | – | 65.1 | not comparable |
| Long context | 66.4 | – | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Dashed: that model has no result on the axis it crosses; the line only joins its neighbours.
Not drawn, no result for these models: Agents & tools, Instruction following, Multimodal.
Where sources disagree
Across the 2 benchmarks both have, Claude Fable 5.2 leads 2, DeepSeek V4 Pro leads 0. Every overall index that rates both puts Claude Fable 5.2 ahead.
| Overall index | Claude Fable 5.2 (max) | DeepSeek V4 Pro (high) | Leader |
|---|---|---|---|
| BenchLeader Index | 69.4 | 59.1 | Claude Fable 5.2 |
| AA Intelligence Index v5.0 | 64.6 | 12.6† | Claude Fable 5.2 |
| Vals Index | 71.3† | 53.0 | Claude Fable 5.2 |
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. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | Claude Fable 5.2 (max) | DeepSeek V4 Pro (high) |
|---|---|---|
| Blended price per 1M tokens | $16.00 | $0.990 |
| Input / output per 1M | $8.00 / $40.00 | $0.660 / $1.98 |
| Cached input per 1M | $0.200 | $0.022 |
| Price per task: chat reply | $0.015 | $0.0009 |
| Price per task: summarize a 30-page report | $0.120 | $0.0091 |
| Price per task: code edit | $0.108 | $0.0069 |
| Price per task: agentic coding session | $0.640 | $0.048 |
| Price per task: structured extraction | $0.024 | $0.0017 |
| Agentic coding session, input mostly cached | $0.289 | $0.019 |
| Cost per benchmark task (AA Index) | $26.54 | – |
| Output tokens per task (AA Index) | 274k | – |
| Of which reasoning | 193k | – |
| Output speed | 72 tok/s | 44 tok/s |
| Time to first answer | 3.0 s | 1.5 s |
Each model is compared at its best-ranked reasoning-effort setting. Claude Fable 5.2 (max) scores 69.4 and takes 3 s to its first answer token; Claude Fable 5.2 (high) scores 68.9 in 3 s, Claude Fable 5.2 (xhigh) scores 68.7 in 3 s. DeepSeek V4 Pro (high) scores 59.1 and takes 1 s to its first answer token; DeepSeek V4 Pro (max) scores 56.9 in 74 s, DeepSeek V4 Pro scores 53.3 in 1 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
132 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 5.2 (max) | DeepSeek V4 Pro (high) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 69.4 | 59.1 |
| Coding score | BenchLeader | 80.2 | 53.8 |
| Knowledge score | BenchLeader | 81.7 | – |
| Long context score | BenchLeader | 66.4 | – |
| Reasoning score | BenchLeader | 78.4 | 62.2 |
| Human preference score | BenchLeader | – | 65.1 |
| Maths score | BenchLeader | – | 64.6 |
| Blended price $/M | BenchLeader | $16.00 | $0.990 |
| Output speed | BenchLeader | 72 tok/s | 44 tok/s |
| Time to first answer | BenchLeader | 3.0 s | 1.5 s |
| Context window | BenchLeader | 1,000,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 90.9% |
| Arena Hard Prompts | Arena (LMArena) | – | 1482 |
| LiveBench 2.0 Reasoning | LiveBench | 88.8%† | – |
| LiveBench 1.0 Reasoning | LiveBench | – | 85.8%† |
| GPQA Diamond (AA) | Artificial Analysis | – | 92.8%† |
| Humanity's Last Exam (AA) | Artificial Analysis | 74.3% | 41.0%† |
| HLE-Diamond (AA) | Artificial Analysis | 80.2% | 19.5%† |
| GPQA Diamond (Vals) | Vals AI | – | 92.4%† |
| Kagi LLM Benchmark | Kagi LLM Benchmark | – | 53.5%† |
| ARC-AGI-1 | ARC Prize | 97.5% | 87.2% |
| ARC-AGI-2 | ARC Prize | 92.5% | 59.7% |
| ARC-AGI-3 | ARC Prize | 94.9% | – |
| CritPt | Artificial Analysis | 35.4% | 18.0%† |
| CritPt v2 hard subset (AA) | Artificial Analysis | 76.5% | 23.5%† |
| MysteryMechanism | Vals AI | 51.8%† | – |
| Chess Puzzles | Epoch AI Benchmarking Hub | – | 16.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | – | 43.0%† |
| CL-bench Life | Epoch AI Benchmarking Hub | – | 13.5% |
| LMCA | Epoch AI Benchmarking Hub | – | 45.5%† |
| DTBench | Epoch AI Benchmarking Hub | – | 93.9%† |
| ForecastBench | ForecastBench | – | 60.7%† |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | – | 77.6%† |
| SciCode | SciCode | – | 46.4% |
| WeirdML | WeirdML | – | 46.5% |
| FrontierCode | Cognition | – | 28.6% |
Show all 132 benchmarks
| Metric | Publisher | Claude Fable 5.2 (max) | DeepSeek V4 Pro (high) |
|---|---|---|---|
| Coding | |||
| Arena Coding | Arena (LMArena) | – | 1503 |
| Arena WebDev | Arena (LMArena) | – | 1581 |
| LiveBench 2.0 Coding | LiveBench | 79.2%† | – |
| LiveBench 1.0 Coding | LiveBench | – | 77.2%† |
| SciCode (AA) | Artificial Analysis | 67.7% | 51.0%† |
| LiveCodeBench | Vals AI | – | 87.5%† |
| IOI | Vals AI | – | 51.6%† |
| SWE-bench (Vals) | Vals AI | – | 96.4%† |
| Code Migration | Vals AI | – | 41.5%† |
| Code Migration v1.1 | Vals AI | 83.3%† | 52.0%† |
| ProgramBench | Vals AI | 19.0%† | 0.0%† |
| SWE-bench Multimodal | Vals AI | 62.3%† | – |
| Vibe Code Bench 1-100 | Vals AI | – | 17.5%† |
| Vibe Code Bench v1.1 | Vals AI | 93.8%† | 82.3%† |
| SWE Atlas: Codebase QnA | Scale AI SEAL | – | 27.1%† |
| SWE Atlas: Test Writing | Scale AI SEAL | – | 27.1%† |
| Surface Evolver Bench | Epoch AI Benchmarking Hub | – | 40.0% |
| ALE-Bench | Epoch AI Benchmarking Hub | – | 1006.1 |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 59.6% | 14.1%† |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | – | 78.7%† |
| Startrise Frontend | Startrise | – | 62.7%† |
| GameBench 2 Overall | DuelLab GameBench 2 | – | 42.0† |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | – | 45.0 |
| Design Arena Website | Design Arena | – | 1231† |
| FrontierCode 1.1 (AA) | Artificial Analysis | 56.2% | 38.8%† |
| FrontierCode 2.0 (AA) | Artificial Analysis | 48.8% | 17.1%† |
| Agents & tools | |||
| APEX-Agents | Mercor | – | 47.3%† |
| LiveBench Long Agentic | LiveBench | 82.9%† | – |
| LiveBench Browser | LiveBench | 86.9%† | – |
| LiveBench Agentic Coding | LiveBench | – | 55.0%† |
| Terminal-Bench 2.1 (Vals) | Vals AI | – | 50.2%† |
| HiL-Bench | Scale AI SEAL | – | 38.5%† |
| GDPval-AA v2.1, normalised score | Artificial Analysis | – | 47.5%† |
| GDPval-AA v2.1 | Artificial Analysis | 1831 | 1457† |
| τ³-Banking (AA) | Artificial Analysis | – | 39.6%† |
| CyberBench | Vals AI | – | 72.9%† |
| Finance Agent v2 | Vals AI | – | 50.4%† |
| Finance Agent v2.1 | Vals AI | 69.5%† | 57.1%† |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 5.0%† | 10.8%† |
| Harvey's Legal Agent Benchmark v1 | Vals AI | – | 7.5%† |
| Legal Research Bench | Vals AI | 46.1%† | 40.9%† |
| SkillsBench | Vals AI | – | 53.8%† |
| SREBench | Vals AI | – | 1.9%† |
| Tax Agent Bench | Vals AI | 32.1%† | 26.8%† |
| Terminal-Bench 2.0 (Vals) | Vals AI | – | 56.2%† |
| Terminal-Bench 4.0 (Vals) | Vals AI | 65.2%† | 14.1%† |
| Terminal-Bench Science | Vals AI | 55.7%† | 4.3%† |
| Vending-Bench 2 | Andon Labs | – | 3284.5† |
| AutomationBench | Artificial Analysis | 76.4% | 56.7%† |
| GDP.pdf (AA) | Artificial Analysis | 24.2% | 11.4%† |
| Harvey LAB-AA v1.1 | Harvey | 7.2% | – |
| EnterpriseOps-Gym | Artificial Analysis | – | 49.6%† |
| AA-Briefcase v1.1 | Artificial Analysis | 1785 | 1253† |
| Design Arena Agentic | Design Arena | – | 921† |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 73.3% | 5.7%† |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | – | 64.6%† |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | – | 26.8%† |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | – | 95.6% |
| ProofBench | Vals AI | – | 50.0%† |
| LiveBench Mathematics | LiveBench | – | 95.1%† |
| AIME 2026 | MathArena | – | 96.7%† |
| HMMT February 2026 | MathArena | – | 93.9%† |
| USAMO 2026 | MathArena | – | 60.7%† |
| MathArena Apex | MathArena | – | 28.1%† |
| Arena Maths | Arena (LMArena) | – | 1465 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | – | 52.9%† |
| LiveBench 2.0 Data Analysis | LiveBench | 81.5%† | – |
| LiveBench Research | LiveBench | 94.5%† | – |
| LiveBench 1.0 Data Analysis | LiveBench | – | 79.2%† |
| AA-Omniscience v1.1 | Artificial Analysis | 66.8 | -6.4† |
| MMLU-Pro | Vals AI | – | 87.3%† |
| LegalBench | Vals AI | – | 82.4%† |
| CorpFin | Vals AI | – | 65.4%† |
| TaxEval | Vals AI | – | 73.1%† |
| BioMysteryBench | Vals AI | 83.7%† | – |
| CaseLaw v2 | Vals AI | – | 59.4%† |
| Excel Modeling Benchmark | Vals AI | – | 52.8%† |
| Excel Modeling Benchmark v1.1 | Vals AI | 77.4%† | 65.1%† |
| MedCode | Vals AI | – | 42.5%† |
| MedScribe | Vals AI | – | 80.2%† |
| Public Benefits Bench v1.1 | Vals AI | – | 62.9%† |
| DrugDiscoveryBench | Scale AI SEAL | – | 31.7%† |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 79.1% | 45.2%† |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 40.7% | 5.9%† |
| HealthBench Professional (AA) | OpenAI | 67.7% | – |
| Instruction following | |||
| LiveBench Language | LiveBench | – | 82.1%† |
| LiveBench Instruction Following | LiveBench | – | 67.7%† |
| Arena Instruction Following | Arena (LMArena) | – | 1461 |
| Human preference | |||
| Arena Text | Arena (LMArena) | – | 1464 |
| EQ-Bench 4 | EQ-Bench | – | 1156† |
| Arena Creative Writing | Arena (LMArena) | – | 1444 |
| Arena Multi-turn | Arena (LMArena) | – | 1469 |
| Arena Longer Queries | Arena (LMArena) | – | 1474 |
| Multimodal | |||
| LiveBench Design | LiveBench | 85.1%† | – |
| AA-3D | Artificial Analysis | 1271 | – |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 83.3% | 80.3%† |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | – | 155.9† |
| LiveBench 2.0 | LiveBench | 84.9%† | – |
| LiveBench 1.0 | LiveBench | – | 77.4%† |
| AA Intelligence Index v5.0 | Artificial Analysis | 64.6 | 23.1† |
| AA Intelligence Index v4.3.2 | Artificial Analysis | – | 36† |
| AA Progress Index v1 | Artificial Analysis | – | 311.5† |
| Vals Index | Vals AI | 71.3† | 53.0† |
| Vals Finance Index | Vals AI | 63.6%† | 52.9%† |
| Vals Science Index | Vals AI | 63.7%† | – |
| Vals Router Index | Vals AI | – | 53.0%† |
| Metadata | |||
| AA Openness Index | Artificial Analysis | – | 44.4† |
Data as of 2026-11-09. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261103-eaea31df94e6 (https://benchmarks.duellab.org/releases/gb2-20261103-eaea31df94e6/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- DeepSeek V4 Pro (high): BenchLeader Index 57.5 → 59.1.
- DeepSeek V4 Pro (high): 1 score revision.
- DeepSeek V4 Pro (high): 1 score revision.
- DeepSeek V4 Pro (high): 16 new results (GPQA Diamond, OTIS Mock AIME, SciCode…); input price moved from $0.44/M to $0.66/M (50% up); output price moved from $0.87/M to $1.98/M (128% up).
- DeepSeek V4 Pro (high): BenchLeader Index 52.7 → 57.5.
- DeepSeek V4 Pro (high): 14 new results (AA Intelligence Index v4.3.2, GPQA Diamond (AA), Humanity's Last Exam (AA)…); input price moved from $0.66/M to $0.44/M (33% down); output price moved from $1.98/M to $0.87/M (56% down).
- DeepSeek V4 Pro (high): BenchLeader Index 55.8 → 52.7.
- DeepSeek V4 Pro (high): 1 new result (Arena WebDev); input price moved from $0.44/M to $0.66/M (50% up); output price moved from $0.87/M to $1.98/M (128% up).
- DeepSeek V4 Pro (high): 1 score revision.
- DeepSeek V4 Pro (high): 1 score revision.
Related comparisons
- Claude Fable 5.2 vs Claude Opus 5.5
- Claude Fable 5.2 vs Gemini 4 Carbon
- Claude Fable 5.2 vs GPT-6 Astra
- DeepSeek V4 Pro vs Claude Opus 5.5
- DeepSeek V4 Pro vs Gemini 4 Carbon
- DeepSeek V4 Pro vs GPT-6 Astra
- Claude Fable 5.2 vs Claude Fable 5.1
- Claude Fable 5.2 vs Claude Fable 5
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Claude Fable 5.2 vs DeepSeek V4 Pro: questions
- Is Claude Fable 5.2 better than DeepSeek V4 Pro?
- Claude Fable 5.2 (max) leads on quality: 69.4 vs 59.1. The BenchLeader Index combines every independent quality benchmark; Claude Fable 5.2 (max) is ahead overall as of 9 Nov 2026, but check the category scores for your use.
- Claude Fable 5.2 vs DeepSeek V4 Pro: which is cheaper, and will that change?
- DeepSeek V4 Pro is cheaper today: $0.990 against $16.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Claude Fable 5.2 or DeepSeek V4 Pro better for coding?
- Claude Fable 5.2 scores higher in coding (80 vs 54 on the category index, where 50 is average).
- Which is faster, Claude Fable 5.2 or DeepSeek V4 Pro?
- Claude Fable 5.2 streams ~1.5× faster (72 vs 44 output tokens per second, Artificial Analysis 7-day median).
- Which is better for long documents?
- Both take about 1M tokens: Claude Fable 5.2 1,000,000, DeepSeek V4 Pro 1,048,576.
More questions (3)
- Which has lower latency?
- DeepSeek V4 Pro answers first: 1.5 s against 3.0 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.
- Do independent benchmarks agree on Claude Fable 5.2 vs DeepSeek V4 Pro?
- Across the 2 benchmarks both have, Claude Fable 5.2 leads 2, DeepSeek V4 Pro leads 0. Every overall index that rates both puts Claude Fable 5.2 ahead.
- When was Claude Fable 5.2 released?
- Claude Fable 5.2 was released on 9 Nov 2026.