GPT-5.1 vs o4-mini
Wins 6 of 7 areas
Coding · Reasoning · Facts · Images and charts · Long documents · Following instructions
Wins 1 of 7 areas
Agents
GPT-5.1 is the stronger all-rounder.o4-mini is cheaper and better at agents.
Scores updated · 36 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.
- ReasoningHard problems that need careful thinking40GPT-5.14 of 4 tests
- FactsGetting facts right instead of making them up40GPT-5.14 of 4 tests
- Images and chartsUnderstanding pictures, charts and video30GPT-5.13 of 3 tests
- Following instructionsDoing exactly what it is asked20GPT-5.12 of 2 tests
- CodingWriting and fixing software21GPT-5.12 of 3 tests
- Long documentsFinding answers in very long texts10GPT-5.11 of 1 test
- AgentsCarrying out multi-step tasks on its own01o4-mini1 of 2 tests · 1 tie
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.
o4-mini costs 51% 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
- Hard command-line tasks in a real terminalTerminal-Bench Hard+30.3points ahead
- Short factual questions, answered correctlySimpleQA Verified+29.2points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+28.6points ahead
Every test, side by side
All 36 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 30.345.515.2+30.3
- SWE-bench VerifiedGPT-5.1 by 8.276.368.1+8.2
- SciCodeo4-mini by 3.243.346.5+3.2
ReasoningGPT-5.1
- Humanity's Last ExamGPT-5.1 by 1228.516.5+12
- ARC-AGI-2GPT-5.1 by 11.517.66.1+11.5
- GPQA DiamondGPT-5.1 by 8.987.378.4+8.9
- CritPtGPT-5.1 by 4.34.90.6+4.3
FactsGPT-5.1
- SimpleQA VerifiedGPT-5.1 by 29.24818.8+29.2
- AA-Omniscience · Non-hallucinationGPT-5.1 by 28.648.119.5+28.6
- AA-Omniscience · AccuracyGPT-5.1 by 12.937.724.8+12.9
- Vectara HHEM hallucination ratelower is betterGPT-5.1 by 6.512.118.6+6.5
Images and chartsGPT-5.1
- MMMU-ProGPT-5.1 by 6.375.569.2+6.3
- LMArena · VisionGPT-5.1 by 56 rating points12501194+56 rating
- MMMUGPT-5.1 by 3.885.481.6+3.8
Following instructionsGPT-5.1
- Multi-ChallengeGPT-5.1 by 20.463.443+20.4
- IFBenchGPT-5.1 by 4.272.968.7+4.2
Other results17 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 41.15.4-35.7+41.1
- τ²-Bench Telecom (AA run)GPT-5.1 by 26.381.955.6+26.3
- Artificial Analysis Coding IndexGPT-5.1 by 23.849.425.6+23.8
- AA Agentic Indexo4-mini by 14.521.636.1+14.5
- ARC-AGI-1GPT-5.1 by 14.172.858.7+14.1
- vectara_avg_summary_lengthGPT-5.1 by 126.7 rating points254.4127.7+126.7 rating
- AA IntelligenceGPT-5.1 by 824.716.7+8
- AIR-Bench 2024GPT-5.1 by 7.786.278.5+7.7
- vectara_factual_consistencyGPT-5.1 by 6.587.981.4+6.5
- frontiermath_tier_4_v1GPT-5.1 by 6.212.56.3+6.2
- bbqo4-mini by 5.388.794+5.3
- AIME 2025GPT-5.1 by 1.39492.7+1.3
- anthropic_red_teamGPT-5.1 by 1.399.598.2+1.3
- vectara_answer_ratetie10099.2tie
- XSTesttie98.297.4tie
- HarmBenchtie97.697tie
- simple_safety_teststie99.8100tie
Questions people ask
Which is better, GPT-5.1 or o4-mini?
GPT-5.1 wins six of the seven areas where both have results: coding, reasoning, facts, images and charts, long documents and following instructions. o4-mini wins agents, and costs 51% less.
Which is better for coding?
GPT-5.1. It wins 2 of the 3 coding tests both models report; o4-mini wins 1.
Which is cheaper?
GPT-5.1 costs $1.25 per million input tokens and $10.00 per million output tokens; o4-mini costs $1.10 and $4.40. That makes o4-mini about 51% cheaper for the same work.
How do you compare the two?
We use the 36 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 17 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.