GPT-5 is the stronger all-rounder.MiniMax M2 is cheaper.
Scores updated · 31 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 software41GPT-54 of 5 tests
- ReasoningHard problems that need careful thinking30GPT-53 of 3 tests
- FactsGetting facts right instead of making them up20GPT-52 of 2 tests
- Long documentsFinding answers in very long texts10GPT-51 of 1 test
- AgentsCarrying out multi-step tasks on its own11Even1 each
- Following instructionsDoing exactly what it is asked00Even0 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.
MiniMax M2 costs 87% 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 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+17.1points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+14.8points ahead
- Reasons across sets of long documentsAA-LCR+13.9points ahead
Where MiniMax M2 pulls ahead
- Real work tasks from 44 professionsGDPVal+5.6points ahead
- Fixes real GitHub issues in many programming languagesSWE-bench Multilingual+1.2points ahead
Every test, side by side
All 31 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5
- LMArena · WebDevGPT-5 by 121 rating points14181297+121 rating
- SciCodeGPT-5 by 6.842.936.1+6.8
- Terminal-Bench HardGPT-5 by 6.832.625.8+6.8
- SWE-bench VerifiedGPT-5 by 5.574.969.4+5.5
- SWE-bench MultilingualMiniMax M2 by 1.255.356.5+1.2
ReasoningGPT-5
- Humanity's Last ExamGPT-5 by 14.828.513.7+14.8
- GPQA DiamondGPT-5 by 7.785.477.7+7.7
- CritPtGPT-5 by 4.85.70.9+4.8
FactsGPT-5
- AA-Omniscience · AccuracyGPT-5 by 17.140.323.2+17.1
- AA-Omniscience · Non-hallucinationGPT-5 by 8.717.89.1+8.7
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 by 37.8-8.7-46.5+37.8
- AA Agentic IndexMiniMax M2 by 2126.547.5+21
- AIME 2025GPT-5 by 16.694.678+16.6
- BrowseComp-ZHGPT-5 by 14.56348.5+14.5
- Artificial Analysis Coding IndexGPT-5 by 8.637.829.2+8.6
- ArtifactsBenchGPT-5 by 6.27366.8+6.2
- xbench-DeepSearchGPT-5 by 5.877.872+5.8
- Terminal-Bench 2.0GPT-5 by 5.235.230+5.2
- MMLU-ProGPT-5 by 5.187.182+5.1
- τ²-BenchMiniMax M2 by 4.682.487+4.6
- AA IntelligenceGPT-5 by 4.42318.6+4.4
- HLE (with tools)GPT-5 by 3.435.231.8+3.4
- Terminal-BenchMiniMax M2 by 2.543.846.3+2.5
- τ²-Bench Telecom (AA run)MiniMax M2 by 284.886.8+2
- FinSearchComp-globalMiniMax M2 by 1.663.965.5+1.6
- GAIA (text only)tie76.475.7tie
- LiveCodeBenchtie83.583tie
Questions people ask
Which is better, GPT-5 or MiniMax M2?
GPT-5 wins four of the six areas where both have results: coding, reasoning, facts and long documents. MiniMax M2 wins none, but costs 87% less. They are level on agents and following instructions.
Which is better for coding?
GPT-5. It wins 4 of the 5 coding tests both models report; MiniMax M2 wins 1.
Which is cheaper?
GPT-5 costs $1.25 per million input tokens and $10.00 per million output tokens; MiniMax M2 costs $0.30 and $1.20. That makes MiniMax M2 about 87% cheaper for the same work.
How do you compare the two?
We use the 31 benchmark tests both models have published scores on. The verdict counts the 14 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.