Claude Fable 5.1 vs Claude Sonnet 5
Wins 8 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Long documents · Following instructions
Wins 0 of 8 areas
—
Claude Fable 5.1 is the stronger all-rounder.Claude Sonnet 5 is cheaper.
Scores updated · 52 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software70Claude Fable 5.17 of 7 tests
- ReasoningHard problems that need careful thinking50Claude Fable 5.15 of 5 tests
- AgentsCarrying out multi-step tasks on its own30Claude Fable 5.13 of 3 tests
- MathCompetition and research-level math30Claude Fable 5.13 of 3 tests
- FactsGetting facts right instead of making them up21Claude Fable 5.12 of 3 tests
- Images and chartsUnderstanding pictures, charts and video10Claude Fable 5.11 of 1 test
- Long documentsFinding answers in very long texts10Claude Fable 5.11 of 1 test
- Following instructionsDoing exactly what it is asked10Claude Fable 5.11 of 1 test
Images and charts, long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
Claude Sonnet 5 costs 80% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Claude Fable 5.1 pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+58.5points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+37.9points ahead
- Short factual questions, answered correctlySimpleQA Verified+37.1points ahead
Where Claude Sonnet 5 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+33.2points ahead
Every test, side by side
All 52 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Fable 5.1
- LMArena · WebDevClaude Fable 5.1 by 200 rating points17411541+200 rating
- SWE-bench ProClaude Fable 5.1 by 1881.263.2+18
- Terminal-Bench 2.1Claude Fable 5.1 by 10.991.480.5+10.9
- SWE-bench MultilingualClaude Fable 5.1 by 10.889.178.3+10.8
- SciCodeClaude Fable 5.1 by 8.863.154.3+8.8
- LiveBench · Agentic CodingClaude Fable 5.1 by 6.766.159.4+6.7
- LiveBench · CodingClaude Fable 5.1 by 5.786.480.7+5.7
AgentsClaude Fable 5.1
- Terminal-Bench 4.0Claude Fable 5.1 by 37.95214.1+37.9
- GDPValClaude Fable 5.1 by 14.562.748.2+14.5
- τ-Bench V3 · BankingClaude Fable 5.1 by 9.947.237.3+9.9
ReasoningClaude Fable 5.1
- SimpleBenchClaude Fable 5.1 by 2686.660.6+26
- Humanity's Last ExamClaude Fable 5.1 by 17.859.141.3+17.8
- CritPtClaude Fable 5.1 by 12.829.716.9+12.8
- LiveBench · ReasoningClaude Fable 5.1 by 391.788.7+3
- GPQA DiamondClaude Fable 5.1 by 2.693.791.1+2.6
FactsClaude Fable 5.1
- SimpleQA VerifiedClaude Fable 5.1 by 37.170.833.7+37.1
- AA-Omniscience · Non-hallucinationClaude Sonnet 5 by 33.227.460.6+33.2
- AA-Omniscience · AccuracyClaude Fable 5.1 by 27.167.240+27.1
Images and chartsClaude Fable 5.1
- LMArena · VisionClaude Fable 5.1 by 44 rating points13181274+44 rating
MathClaude Fable 5.1
- FrontierMath Tier 4Claude Fable 5.1 by 58.587.829.3+58.5
- FrontierMath Tiers 1-3 (v2)Claude Fable 5.1 by 24.690.265.6+24.6
- LiveBench · MathematicsClaude Fable 5.1 by 4.19792.9+4.1
Following instructionsClaude Fable 5.1
- LiveBench · Instruction FollowingClaude Fable 5.1 by 9.17363.9+9.1
Other results28 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- Terminal-Bench 4.0Claude Fable 5.1 by 43.455.812.4+43.4
- OSWorld 2.0 (partial)Claude Fable 5.1 by 35.377.942.6+35.3
- GDPval-AA 2.1Claude Fable 5.1 by 286 rating points17351449+286 rating
- AA-OmniscienceClaude Fable 5.1 by 2743.516.4+27
- SWE-bench MultimodalClaude Fable 5.1 by 26.654.728.1+26.6
- GDPval-AA v2 EloClaude Fable 5.1 by 251 rating points18351584+251 rating
- OSWorld 2.1 partialClaude Fable 5.1 by 23.780.757+23.7
- GDPval-AA v2Claude Fable 5.1 by 235 rating points18531618+235 rating
- AA IntelligenceClaude Fable 5.1 by 15.253.438.2+15.2
- livebench_languageClaude Fable 5.1 by 14.589.575+14.5
- AA Agentic IndexClaude Fable 5.1 by 13.75844.3+13.7
- DeepSWE 1.1Claude Fable 5.1 by 13.467.454+13.4
- HealthBench Professional (raw score)Claude Fable 5.1 by 11.874.262.4+11.8
- HealthBench Professional (raw)Claude Fable 5.1 by 11.874.262.4+11.8
- LVBenchClaude Fable 5.1 by 11.279.768.5+11.2
- Program BenchClaude Fable 5.1 by 10.387.677.3+10.3
- Artificial Analysis Coding IndexClaude Fable 5.1 by 10.181.671.5+10.1
- OfficeQA ProClaude Fable 5.1 by 9.66959.4+9.6
- livebench_data_analysisClaude Fable 5.1 by 8.680.371.7+8.6
- HLE (with tools)Claude Fable 5.1 by 7.66557.4+7.6
- HealthBench (raw score)Claude Fable 5.1 by 7.566.759.2+7.5
- HealthBench (raw)Claude Fable 5.1 by 7.566.759.2+7.5
- OfficeQAClaude Fable 5.1 by 5.180.275.1+5.1
- GMMLUClaude Fable 5.1 by 4.89489.2+4.8
- HealthBench ProfessionalClaude Fable 5.1 by 4.362.157.8+4.3
- MILUClaude Fable 5.1 by 3.79389.3+3.7
- GDP (Surge AI)Claude Fable 5.1 by 3.585.181.6+3.5
- Toolathlon VerifiedClaude Fable 5.1 by 3.177.874.7+3.1
Questions people ask
Which is better, Claude Fable 5.1 or Claude Sonnet 5?
Claude Fable 5.1 wins all eight areas we test: coding, agents, reasoning, facts, images and charts, math, long documents and following instructions. Claude Sonnet 5 wins none, but costs 80% less.
Which is better for coding?
Claude Fable 5.1. It wins 7 of the 7 coding tests both models report; Claude Sonnet 5 wins none.
Which is cheaper?
Claude Fable 5.1 costs $10.00 per million input tokens and $50.00 per million output tokens; Claude Sonnet 5 costs $2.00 and $10.00. That makes Claude Sonnet 5 about 80% cheaper for the same work.
How do you compare the two?
We use the 52 benchmark tests both models have published scores on. The verdict counts the 24 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 28 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.