GPT-5.4 mini vs GPT-5.6 Sol
Wins 0 of 8 areas
—
Wins 7 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents · Following instructions
GPT-5.6 Sol is the stronger all-rounder.GPT-5.4 mini is cheaper.
Scores updated · 43 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 software07GPT-5.6 Sol7 of 7 tests
- AgentsCarrying out multi-step tasks on its own07GPT-5.6 Sol7 of 7 tests
- ReasoningHard problems that need careful thinking05GPT-5.6 Sol5 of 5 tests
- MathCompetition and research-level math03GPT-5.6 Sol3 of 3 tests
- Images and chartsUnderstanding pictures, charts and video02GPT-5.6 Sol2 of 2 tests
- Following instructionsDoing exactly what it is asked01GPT-5.6 Sol1 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts01GPT-5.6 Sol1 of 1 test
- FactsGetting facts right instead of making them up22Even2 each
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-5.4 mini costs 78% 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-5.6 Sol pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+73.6points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+73.2points ahead
- Short factual questions, answered correctlySimpleQA Verified+40.3points ahead
Where GPT-5.4 mini pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+2points ahead
Every test, side by side
All 43 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.6 Sol
- Terminal-Bench 2.1GPT-5.6 Sol by 28.859.288+28.8
- LMArena · WebDevGPT-5.6 Sol by 221 rating points13971618+221 rating
- LiveBench · Agentic CodingGPT-5.6 Sol by 14.541.756.2+14.5
- Terminal-Bench HardGPT-5.6 Sol by 13.652.365.9+13.6
- LiveBench · CodingGPT-5.6 Sol by 12.371.683.9+12.3
- SWE-bench ProGPT-5.6 Sol by 10.254.464.6+10.2
- SciCodeGPT-5.6 Sol by 552.157.1+5
AgentsGPT-5.6 Sol
- Terminal-Bench 4.0GPT-5.6 Sol by 37.9239.9+37.9
- GDPValGPT-5.6 Sol by 29.825.855.6+29.8
- MCP AtlasGPT-5.6 Sol by 25.957.783.6+25.9
- AA IT-Bench SREGPT-5.6 Sol by 2135.256.2+21
- τ-Bench V3 · BankingGPT-5.6 Sol by 18.725.644.3+18.7
- AA ApexAgentsGPT-5.6 Sol by 11.728.239.9+11.7
- OSWorld-VerifiedGPT-5.6 Sol by 10.972.183+10.9
ReasoningGPT-5.6 Sol
- ARC-AGI-2GPT-5.6 Sol by 73.618.992.5+73.6
- CritPtGPT-5.6 Sol by 22.31032.3+22.3
- Humanity's Last ExamGPT-5.6 Sol by 21.428.149.5+21.4
- LiveBench · ReasoningGPT-5.6 Sol by 20.471.391.7+20.4
- GPQA DiamondGPT-5.6 Sol by 6.687.594.1+6.6
FactsEven
- SimpleQA VerifiedGPT-5.6 Sol by 40.329.469.7+40.3
- AA-Omniscience · AccuracyGPT-5.6 Sol by 21.937.559.4+21.9
- Vectara HHEM hallucination ratelower is betterGPT-5.4 mini by 6.95.512.4+6.9
- AA-Omniscience · Non-hallucinationGPT-5.4 mini by 29.87.8+2
Images and chartsGPT-5.6 Sol
- MMMU-ProGPT-5.6 Sol by 10.173.383.4+10.1
- LMArena · VisionGPT-5.6 Sol by 34 rating points12451279+34 rating
MathGPT-5.6 Sol
- FrontierMath Tier 4GPT-5.6 Sol by 73.29.883+73.2
- FrontierMath Tiers 1-3 (v2)GPT-5.6 Sol by 37.951.289.1+37.9
- LiveBench · MathematicsGPT-5.6 Sol by 17.778.596.2+17.7
Following instructionsGPT-5.6 Sol
- LiveBench · Instruction FollowingGPT-5.6 Sol by 12.159.871.8+12.1
- IFBenchtie73.372.7tie
Other results12 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-5.6 Sol by 40.9-18.922+40.9
- ARC-AGI-1GPT-5.6 Sol by 32.863.796.5+32.8
- AA Agentic IndexGPT-5.6 Sol by 30.819.750.5+30.8
- AA IntelligenceGPT-5.6 Sol by 22.924.147+22.9
- Artificial Analysis Coding IndexGPT-5.6 Sol by 21.356.177.4+21.3
- livebench_languageGPT-5.6 Sol by 16.77187.7+16.7
- ToolathlonGPT-5.6 Sol by 15.142.958+15.1
- livebench_data_analysisGPT-5.6 Sol by 970.879.8+9
- vectara_avg_summary_lengthGPT-5.6 Sol by 76.4 rating points54.7131.1+76.4 rating
- vectara_factual_consistencyGPT-5.4 mini by 6.994.587.6+6.9
- τ²-Bench Telecom (AA run)GPT-5.6 Sol by 1.883.385.1+1.8
- vectara_answer_ratetie10099.1tie
Questions people ask
Which is better, GPT-5.4 mini or GPT-5.6 Sol?
GPT-5.6 Sol wins seven of the eight areas we test: coding, agents, reasoning, images and charts, math, long documents and following instructions. GPT-5.4 mini wins none, but costs 78% less. They are level on facts.
Which is better for coding?
GPT-5.6 Sol. It wins 7 of the 7 coding tests both models report; GPT-5.4 mini wins none.
Which is cheaper?
GPT-5.4 mini costs $0.75 per million input tokens and $4.50 per million output tokens; GPT-5.6 Sol costs $4.00 and $20.00. That makes GPT-5.4 mini about 78% cheaper for the same work.
How do you compare the two?
We use the 43 benchmark tests both models have published scores on. The verdict counts the 31 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 12 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.