DeepSeek-V3.2 vs GPT-5.1
Wins 0 of 6 areas
—
Wins 5 of 6 areas
Coding · Reasoning · Facts · Long documents · Following instructions
GPT-5.1 is the stronger all-rounder.DeepSeek-V3.2 is cheaper.
Scores updated · 37 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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 software06GPT-5.16 of 6 tests
- ReasoningHard problems that need careful thinking04GPT-5.14 of 4 tests
- FactsGetting facts right instead of making them up13GPT-5.13 of 4 tests
- Long documentsFinding answers in very long texts01GPT-5.11 of 1 test
- Following instructionsDoing exactly what it is asked01GPT-5.11 of 1 test
- AgentsCarrying out multi-step tasks on its own11Even1 each · 1 tie
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.
DeepSeek-V3.2 costs 94% 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
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+30.8points ahead
- Short factual questions, answered correctlySimpleQA Verified+20.5points ahead
- Abstract visual puzzles that people can solveARC-AGI-2+13.6points ahead
Where DeepSeek-V3.2 pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+2.9points ahead
Every test, side by side
All 37 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.1
- Terminal-Bench HardGPT-5.1 by 9.935.645.5+9.9
- Terminal-Bench 2.1GPT-5.1 by 5.646.852.4+5.6
- SciCodeGPT-5.1 by 4.438.943.3+4.4
- LiveCodeBench v6GPT-5.1 by 3.783.387+3.7
- LMArena · WebDevGPT-5.1 by 33 rating points13621395+33 rating
- SWE-bench VerifiedGPT-5.1 by 3.273.176.3+3.2
AgentsEven
- GDPValGPT-5.1 by 6.79.816.5+6.7
- τ-Bench V3 · BankingDeepSeek-V3.2 by 2.918.815.9+2.9
- BrowseComptie51.450.8tie
ReasoningGPT-5.1
- ARC-AGI-2GPT-5.1 by 13.6417.6+13.6
- Humanity's Last ExamGPT-5.1 by 3.924.628.5+3.9
- GPQA DiamondGPT-5.1 by 3.38487.3+3.3
- CritPtGPT-5.1 by 22.94.9+2
FactsGPT-5.1
- AA-Omniscience · Non-hallucinationGPT-5.1 by 30.817.348.1+30.8
- SimpleQA VerifiedGPT-5.1 by 20.527.548+20.5
- Vectara HHEM hallucination ratelower is betterDeepSeek-V3.2 by 5.86.312.1+5.8
- AA-Omniscience · AccuracyGPT-5.1 by 4.73337.7+4.7
Other results18 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.
- AA-OmniscienceGPT-5.1 by 27.9-22.55.4+27.9
- vectara_avg_summary_lengthGPT-5.1 by 192.4 rating points62254.4+192.4 rating
- ARC-AGI-1GPT-5.1 by 15.85772.8+15.8
- frontiermath_tier_4_v1GPT-5.1 by 10.42.112.5+10.4
- τ²-Bench Telecom (AA run)DeepSeek-V3.2 by 8.790.681.9+8.7
- vectara_answer_rateGPT-5.1 by 7.492.6100+7.4
- swe_bench_bashGPT-5.1 by 66066+6
- vectara_factual_consistencyDeepSeek-V3.2 by 5.893.787.9+5.8
- Artificial Analysis Coding IndexGPT-5.1 by 5.244.249.4+5.2
- HMMT Feb. 2025GPT-5.1 by 3.892.596.3+3.8
- AA Agentic IndexGPT-5.1 by 3.318.321.6+3.3
- AA IntelligenceGPT-5.1 by 3.221.524.7+3.2
- τ²-BenchGPT-5.1 by 2.480.382.7+2.4
- MMLU-ProGPT-5.1 by 28587+2
- HLE (with tools)GPT-5.1 by 1.940.842.7+1.9
- HMMT Nov. 2025GPT-5.1 by 1.79091.7+1.7
- Terminal-Bench 2.0GPT-5.1 by 1.246.447.6+1.2
- AIME 2025tie93.194tie
Questions people ask
Which is better, DeepSeek-V3.2 or GPT-5.1?
GPT-5.1 wins five of the six areas where both have results: coding, reasoning, facts, long documents and following instructions. DeepSeek-V3.2 wins none, but costs 94% less. They are level on agents.
Which is better for coding?
GPT-5.1. It wins 6 of the 6 coding tests both models report; DeepSeek-V3.2 wins none.
Which is cheaper?
DeepSeek-V3.2 costs $0.28 per million input tokens and $0.42 per million output tokens; GPT-5.1 costs $1.25 and $10.00. That makes DeepSeek-V3.2 about 94% cheaper for the same work.
How do you compare the two?
We use the 37 benchmark tests both models have published scores on. The verdict counts the 19 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 18 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.