Claude Sonnet 4 vs GPT-5.1
Wins 1 of 7 areas
Facts
Wins 6 of 7 areas
Coding · Agents · Reasoning · Images and charts · Long documents · Following instructions
GPT-5.1 is the stronger all-rounder.Claude Sonnet 4 is better at facts.
Scores updated · 45 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. 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 software05GPT-5.15 of 5 tests
- ReasoningHard problems that need careful thinking05GPT-5.15 of 5 tests
- Images and chartsUnderstanding pictures, charts and video03GPT-5.13 of 3 tests
- AgentsCarrying out multi-step tasks on its own02GPT-5.12 of 3 tests · 1 tie
- Following instructionsDoing exactly what it is asked02GPT-5.12 of 2 tests
- Long documentsFinding answers in very long texts01GPT-5.11 of 1 test
- FactsGetting facts right instead of making them up21Claude Sonnet 42 of 3 tests
Long documents rests on a single test.
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.
GPT-5.1 costs 38% 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 GPT-5.1 pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+38.6points ahead
- Recent programming contest problemsLiveCodeBench v6+38.5points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+18.2points ahead
Where Claude Sonnet 4 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+22.8points ahead
Every test, side by side
All 45 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.1
- LiveCodeBench v6GPT-5.1 by 38.548.587+38.5
- Terminal-Bench 2.1GPT-5.1 by 16.136.352.4+16.1
- Terminal-Bench HardGPT-5.1 by 14.431.145.5+14.4
- SWE-bench VerifiedGPT-5.1 by 3.672.776.3+3.6
- SciCodeGPT-5.1 by 3.34043.3+3.3
AgentsGPT-5.1
- BrowseCompGPT-5.1 by 38.612.250.8+38.6
- GDPValGPT-5.1 by 6.89.716.5+6.8
- τ-Bench V3 · Bankingtie16.715.9tie
ReasoningGPT-5.1
- Humanity's Last ExamGPT-5.1 by 17.810.728.5+17.8
- ARC-AGI-2GPT-5.1 by 11.75.917.6+11.7
- GPQA DiamondGPT-5.1 by 9.677.787.3+9.6
- SimpleBenchGPT-5.1 by 7.745.553.2+7.7
- CritPtGPT-5.1 by 4.60.34.9+4.6
FactsClaude Sonnet 4
- AA-Omniscience · Non-hallucinationClaude Sonnet 4 by 22.870.948.1+22.8
- AA-Omniscience · AccuracyGPT-5.1 by 1522.737.7+15
- Vectara HHEM hallucination ratelower is betterClaude Sonnet 4 by 1.810.312.1+1.8
Images and chartsGPT-5.1
- MMMU-ProGPT-5.1 by 13.761.875.5+13.7
- MMMUGPT-5.1 by 1174.485.4+11
- LMArena · VisionGPT-5.1 by 59 rating points11911250+59 rating
Following instructionsGPT-5.1
- IFBenchGPT-5.1 by 18.254.772.9+18.2
- Multi-ChallengeGPT-5.1 by 6.357.163.4+6.3
Other results23 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.
- τ²-BenchGPT-5.1 by 37.545.282.7+37.5
- ARC-AGI-1GPT-5.1 by 32.84072.8+32.8
- AIME 2025GPT-5.1 by 23.570.594+23.5
- HLE (with tools)GPT-5.1 by 22.420.342.7+22.4
- τ²-Bench Telecom (AA run)GPT-5.1 by 17.364.681.9+17.3
- frontiermath_tier_4_v1GPT-5.1 by 12.5012.5+12.5
- Artificial Analysis Coding IndexGPT-5.1 by 11.837.649.4+11.8
- Tau2 airlineGPT-5.1 by 11.555.567+11.5
- vectara_avg_summary_lengthGPT-5.1 by 108.6 rating points145.8254.4+108.6 rating
- bbqClaude Sonnet 4 by 7.996.688.7+7.9
- AA IntelligenceGPT-5.1 by 5.818.924.7+5.8
- AA-OmniscienceGPT-5.1 by 5.20.25.4+5.2
- MMMLUGPT-5.1 by 4.586.591+4.5
- AA Agentic IndexGPT-5.1 by 417.621.6+4
- MMLU-ProGPT-5.1 by 3.383.787+3.3
- τ²-Bench (Retail)GPT-5.1 by 2.97577.9+2.9
- AIR-Bench 2024Claude Sonnet 4 by 2.188.386.2+2.1
- vectara_factual_consistencyClaude Sonnet 4 by 1.889.787.9+1.8
- XSTestGPT-5.1 by 1.696.698.2+1.6
- vectara_answer_rateGPT-5.1 by 1.498.6100+1.4
- HarmBenchtie9897.6tie
- anthropic_red_teamtie99.299.5tie
- simple_safety_teststie10099.8tie
Questions people ask
Which is better, Claude Sonnet 4 or GPT-5.1?
GPT-5.1 wins six of the seven areas where both have results: coding, agents, reasoning, images and charts, long documents and following instructions. Claude Sonnet 4 wins facts.
Which is better for coding?
GPT-5.1. It wins 5 of the 5 coding tests both models report; Claude Sonnet 4 wins none.
Which is cheaper?
Claude Sonnet 4 costs $3.00 per million input tokens and $15.00 per million output tokens; GPT-5.1 costs $1.25 and $10.00. That makes GPT-5.1 about 38% cheaper for the same work.
How do you compare the two?
We use the 45 benchmark tests both models have published scores on. The verdict counts the 22 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 23 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.