LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Fable 5 vs DeepSeek V4 Pro 0813

Updated October 21, 2026. 71 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Claude Fable 5 Replace DeepSeek V4 Pro 0813
Claude Fable 5Anthropic
Coding
82.1
Overall
49.6
Output $/M
$50.00
68
DeepSeek V4 Pro 0813DeepSeek
Coding
54.6
Overall
36.0
Output $/M
$1.98
3
vs boards won of 71 shared

Decision reading

Pick Claude Fable 5 for coding: it scores 82.1 on the Vibe Coding Index against 54.6 for DeepSeek V4 Pro 0813.

Pick DeepSeek V4 Pro 0813 when output tokens cost 25.3x less, or when a solved task costs $0.10 against $2.16.

Model AClaude Fable 5Anthropic82.1/1002 independent benchmarks$2.16 per solved taskModel BDeepSeek V4 Pro 0813DeepSeek54.6/1001 independent benchmark$0.10 per solved task
Shared results
71
Claude Fable 5 only
55
DeepSeek V4 Pro 0813 only
7
Like-for-like categories
7 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Claude Fable 5 82.1 from 2 benchmarks, DeepSeek V4 Pro 0813 54.6 from 1 benchmark.
Widest category
Math, 26.2 points apart on average over 6 shared benchmarks.
Widest single gap
FrontierMath, tier 4: Claude Fable 5 90.2% against DeepSeek V4 Pro 0813 26.8%, 63.4 points apart on a 0 to 100 scale.
Output price
Claude Fable 5 $50.00 against DeepSeek V4 Pro 0813 $1.98 per million output tokens, 25.3x apart.
Context window
Claude Fable 5 takes 1M tokens, DeepSeek V4 Pro 0813 1.05M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryClaude Fable 5BarsDeepSeek V4 Pro 0813Ties
Coding161
Intelligence20
Agentic work70
Reasoning92
Math60
Knowledge100
General180
All shared boards683
CodingLike-for-likeFour of these boards feed the Vibe Coding Index161
Coding benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
LiveBench coding98.3%Not measuredOnly Claude Fable 5 measured
LMArena WebDev16241582Claude Fable 5
AA Coding Index76.568.8Claude Fable 5
LiveCodeBench (Vals AI)89.8%87.5%Claude Fable 5
SWE-bench Verified (Vals AI)95.0%96.4%DeepSeek V4 Pro 0813
Terminal-Bench 4 (Vals AI)41.4%14.1%Claude Fable 5
Vibe Code Bench (Vals AI)90.3%82.3%Claude Fable 5
Code migration (Vals AI)55.1%41.5%Claude Fable 5
ProgramBench (Vals AI)2.0%0.0%Claude Fable 5
IOI (Vals AI)Not measured51.6%Only DeepSeek V4 Pro 0813 measured
Terminal-Bench 2.1 (Vals AI)80.5%54.7%Claude Fable 5
Vibe Code Bench 1-100 (Vals AI)Not measured17.5%Only DeepSeek V4 Pro 0813 measured
SkillsBench (Vals AI)Not measured53.8%Only DeepSeek V4 Pro 0813 measured
CyberBench (Vals AI)Not measured72.9%Only DeepSeek V4 Pro 0813 measured
SciCode61.0%51.0%Claude Fable 5
Terminal-Bench Hard (Artificial Analysis run)62.9%Not measuredOnly Claude Fable 5 measured
FrontierCode53.5%28.5%Claude Fable 5
FrontierSWE47.0%Not measuredOnly Claude Fable 5 measured
GSO-Bench78.4%Not measuredOnly Claude Fable 5 measured
WeirdML91.9%66.2%Claude Fable 5
MirrorCode63.9%Not measuredOnly Claude Fable 5 measured
ALE-Bench2041.31403.2Claude Fable 5
LMArena Text: coding15181469Claude Fable 5
LMArena Text: software and IT15131462Claude Fable 5
LMArena WebDev: React16311588Claude Fable 5
LMArena WebDev: HTML16511566Claude Fable 5
LMArena WebDev: from an image1623Not measuredOnly Claude Fable 5 measured
Neither model has a result on the other 11 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
AA Intelligence Index49.636.0Claude Fable 5
Epoch Capabilities Index163.4155.6Claude Fable 5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index70
Agentic work benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
AA Agentic Index50.741.3Claude Fable 5
Design Arena: agents building web apps1261Not measuredOnly Claude Fable 5 measured
Design Arena: agents building full-stack apps1280Not measuredOnly Claude Fable 5 measured
Design Arena: agents building mobile apps1264Not measuredOnly Claude Fable 5 measured
Design Arena: agents building Android apps1289Not measuredOnly Claude Fable 5 measured
Design Arena: agents building Godot games1292Not measuredOnly Claude Fable 5 measured
Design Arena: agents building games1285Not measuredOnly Claude Fable 5 measured
Design Arena: HTML slides1277Not measuredOnly Claude Fable 5 measured
Design Arena: PowerPoint slides1295Not measuredOnly Claude Fable 5 measured
LiveBench research86.9%Not measuredOnly Claude Fable 5 measured
LiveBench browser92.0%Not measuredOnly Claude Fable 5 measured
LiveBench long agentic53.2%Not measuredOnly Claude Fable 5 measured
SREBench (Vals AI)Not measured1.9%Only DeepSeek V4 Pro 0813 measured
Finance Agent (Vals AI)56.3%50.4%Claude Fable 5
Tax Agent Bench (Vals AI)38.3%26.8%Claude Fable 5
Legal Research Bench (Vals AI)49.5%40.9%Claude Fable 5
Harvey's Legal Agent Benchmark (Vals AI)Not measured10.8%Only DeepSeek V4 Pro 0813 measured
Excel Modeling Benchmark (Vals AI)73.7%52.8%Claude Fable 5
Terminal-Bench-Science (Vals AI)15.7%4.3%Claude Fable 5
τ²-Bench Telecom98.5%Not measuredOnly Claude Fable 5 measured
APEX-Agents63.6%47.3%Claude Fable 5
Remote Labor Index16.1%Not measuredOnly Claude Fable 5 measured
PostTrainBench41.8%Not measuredOnly Claude Fable 5 measured
Vending-Bench 2$5,680Not measuredOnly Claude Fable 5 measured
LMArena Search1229Not measuredOnly Claude Fable 5 measured
Neither model has a result on the other 14 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index92
Reasoning benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
GPQA Diamond93.2%92.4%Claude Fable 5
LiveBench reasoning89.7%Not measuredOnly Claude Fable 5 measured
MMMU Pro (Vals AI)89.3%Not measuredOnly Claude Fable 5 measured
GPQA Diamond (Artificial Analysis run)92.6%92.8%DeepSeek V4 Pro 0813
Humanity's Last Exam (Artificial Analysis run)55.5%41.0%Claude Fable 5
GPQA Diamond (Epoch AI run)85.9%91.7%DeepSeek V4 Pro 0813
ARC-AGI98.5%90.5%Claude Fable 5
ARC-AGI-289.2%61.3%Claude Fable 5
SimpleBench81.9%Not measuredOnly Claude Fable 5 measured
Chess puzzles47.0%46.0%Claude Fable 5
Mystery game puzzles52.0%43.0%Claude Fable 5
EBR-bench39.5%Not measuredOnly Claude Fable 5 measured
LMArena Text: hard prompts15071458Claude Fable 5
LMArena Text: expert15341459Claude Fable 5
LMArena Text: hard prompts in English15091463Claude Fable 5
LMArena Vision: diagrams1357Not measuredOnly Claude Fable 5 measured
LMArena Vision: homework1361Not measuredOnly Claude Fable 5 measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index60
Math benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
OTIS Mock AIME100.0%98.6%Claude Fable 5
FrontierMath, tiers 1 to 387.0%64.6%Claude Fable 5
FrontierMath, tier 490.2%26.8%Claude Fable 5
ProofBench95.0%50.0%Claude Fable 5
LMArena Text: math15161441Claude Fable 5
LMArena Text: mathematical work15151443Claude Fable 5
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index100
Knowledge benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
MMLU Pro91.5%87.0%Claude Fable 5
LegalBench (Vals AI)88.6%82.4%Claude Fable 5
CorpFin (Vals AI)71.8%65.4%Claude Fable 5
TaxEval (Vals AI)76.9%73.1%Claude Fable 5
MortgageTax (Vals AI)68.9%Not measuredOnly Claude Fable 5 measured
MedCode (Vals AI)56.1%42.5%Claude Fable 5
SimpleQA Verified70.7%52.9%Claude Fable 5
LMArena Text: science15061457Claude Fable 5
LMArena Text: medicine and healthcare14801446Claude Fable 5
LMArena Text: legal and government14981459Claude Fable 5
LMArena Text: business and finance14801442Claude Fable 5
Neither model has a result on the other 6 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
Design Arena: websites1296Not measuredOnly Claude Fable 5 measured
Design Arena: UI components1318Not measuredOnly Claude Fable 5 measured
Design Arena: game dev1336Not measuredOnly Claude Fable 5 measured
Design Arena: data viz1304Not measuredOnly Claude Fable 5 measured
Design Arena: 3D1319Not measuredOnly Claude Fable 5 measured
Design Arena: all code categories1310Not measuredOnly Claude Fable 5 measured
Design Arena: SVG1302Not measuredOnly Claude Fable 5 measured
Design Arena: ASCII art1328Not measuredOnly Claude Fable 5 measured
LiveBench design84.7%Not measuredOnly Claude Fable 5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index180
General benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
Vals IndexNot measured53.0%Only DeepSeek V4 Pro 0813 measured
LiveBench data analysis80.5%Not measuredOnly Claude Fable 5 measured
Public Benefits Bench (Vals AI)70.4%Not measuredOnly Claude Fable 5 measured
SAGE (Vals AI)51.9%Not measuredOnly Claude Fable 5 measured
MedScribe (Vals AI)88.5%80.2%Claude Fable 5
Vals Multimodal Index74.2%Not measuredOnly Claude Fable 5 measured
IFBench63.5%Not measuredOnly Claude Fable 5 measured
Long context reasoning (AA-LCR)82.3%80.3%Claude Fable 5
DTBench98.4%93.9%Claude Fable 5
LMCA61.2%45.5%Claude Fable 5
Surface Evolver Bench95.0%Not measuredOnly Claude Fable 5 measured
Furniture assembly35.8%Not measuredOnly Claude Fable 5 measured
LMArena Text14911446Claude Fable 5
LMArena Text: instruction following15021448Claude Fable 5
LMArena Text: creative writing14941439Claude Fable 5
LMArena Vision1325Not measuredOnly Claude Fable 5 measured
LMArena Text: English14981457Claude Fable 5
LMArena Text: non-English14821431Claude Fable 5
LMArena Text: Chinese15431481Claude Fable 5
LMArena Text: German1486Not measuredOnly Claude Fable 5 measured
LMArena Text: Japanese1504Not measuredOnly Claude Fable 5 measured
LMArena Text: Korean14871449Claude Fable 5
LMArena Text: French15051446Claude Fable 5
LMArena Text: Spanish14981435Claude Fable 5
LMArena Text: Russian15031448Claude Fable 5
LMArena Text: Polish1494Not measuredOnly Claude Fable 5 measured
LMArena Text: multi-turn15041454Claude Fable 5
LMArena Text: longer queries15101458Claude Fable 5
LMArena Text: writing and language15011443Claude Fable 5
LMArena Text: entertainment and media14821424Claude Fable 5
LMArena Vision: English1323Not measuredOnly Claude Fable 5 measured
LMArena Vision: Chinese1373Not measuredOnly Claude Fable 5 measured
LMArena Vision: reading text1337Not measuredOnly Claude Fable 5 measured
LMArena Vision: humor1341Not measuredOnly Claude Fable 5 measured
LMArena Vision: creative writing1338Not measuredOnly Claude Fable 5 measured
LMArena Document1496Not measuredOnly Claude Fable 5 measured
Neither model has a result on the other 10 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Fable 5 and DeepSeek V4 Pro 0813
BenchmarkClaude Fable 5DeepSeek V4 Pro 0813Leader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Claude Fable 5$0.035Fits in one request
DeepSeek V4 Pro 0813$0.0016Fits in one request

DeepSeek V4 Pro 0813 costs less

Repository review

50K input + 3K output tokens

Claude Fable 5$0.650Fits in one request
DeepSeek V4 Pro 0813$0.039Fits in one request

DeepSeek V4 Pro 0813 costs less

Agent loop

200K input + 10K output tokens

Claude Fable 5$2.50Fits in one request
DeepSeek V4 Pro 0813$0.152Fits in one request

DeepSeek V4 Pro 0813 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Claude Fable 5 and DeepSeek V4 Pro 0813
SpecificationAnthropic Claude Fable 5DeepSeek V4 Pro 0813
MakerAnthropicDeepSeek
WeightsNo public weights listedOpen weights
ReleasedJune 9, 2026August 12, 2026
Context window1M1.05M
Input $/Mtok$10.00$0.66
Output $/Mtok$50.00$1.98
Coding index82.154.6
Independent measurements behind it21
Cost per solved task$2.16$0.10
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Fable 5 better than DeepSeek V4 Pro 0813 for coding?

Pick Claude Fable 5 for coding: it scores 82.1 on the Vibe Coding Index against 54.6 for DeepSeek V4 Pro 0813. Pick DeepSeek V4 Pro 0813 when output tokens cost 25.3x less, or when a solved task costs $0.10 against $2.16.

Which is cheaper, Claude Fable 5 or DeepSeek V4 Pro 0813?

Claude Fable 5 costs $10.00 per million input tokens and $50.00 per million output tokens; DeepSeek V4 Pro 0813 costs $0.66 and $1.98. Per solved SWE-bench Verified task, Claude Fable 5 costs $2.16 and DeepSeek V4 Pro 0813 $0.10.

Which has the bigger context window, Claude Fable 5 or DeepSeek V4 Pro 0813?

DeepSeek V4 Pro 0813, with 1.05M tokens against 1M.

How many benchmarks is the Claude Fable 5 vs DeepSeek V4 Pro 0813 comparison based on?

71 benchmarks have a result for both models as of October 21, 2026. Claude Fable 5's coding index rests on 2 independent benchmarks and DeepSeek V4 Pro 0813's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next