GPT-5 vs MiniMax M2.5
Wins 4 of 6 areas
Reasoning · Facts · Long documents · Following instructions
Wins 2 of 6 areas
Coding · Agents
GPT-5 is the stronger all-rounder.MiniMax M2.5 is cheaper and better at coding.
Scores updated · 27 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.
- ReasoningHard problems that need careful thinking30GPT-53 of 4 tests · 1 tie
- FactsGetting facts right instead of making them up21GPT-52 of 3 tests
- Long documentsFinding answers in very long texts10GPT-51 of 1 test
- Following instructionsDoing exactly what it is asked10GPT-51 of 1 test
- CodingWriting and fixing software13MiniMax M2.53 of 5 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own02MiniMax M2.52 of 2 tests
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 M2.5 costs 87% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GPT-5 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+14.1points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+8points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+5.9points ahead
Where MiniMax M2.5 pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+21.4points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+13.6points ahead
- Real work tasks from 44 professionsGDPVal+12.9points ahead
Every test, side by side
All 27 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingMiniMax M2.5
- SWE-bench ProMiniMax M2.5 by 13.641.855.4+13.6
- SWE-bench VerifiedMiniMax M2.5 by 5.374.980.2+5.3
- LMArena · WebDevGPT-5 by 31 rating points14181387+31 rating
- Terminal-Bench HardMiniMax M2.5 by 2.232.634.8+2.2
- SciCodetie42.942.6tie
AgentsMiniMax M2.5
- BrowseCompMiniMax M2.5 by 21.454.976.3+21.4
- GDPValMiniMax M2.5 by 12.92133.9+12.9
ReasoningGPT-5
- Humanity's Last ExamGPT-5 by 828.520.5+8
- ARC-AGI-2GPT-5 by 59.94.9+5
- CritPtGPT-5 by 4.65.71.1+4.6
- GPQA Diamondtie85.484.8tie
FactsGPT-5
- AA-Omniscience · AccuracyGPT-5 by 14.140.326.2+14.1
- Vectara HHEM hallucination ratelower is betterMiniMax M2.5 by 615.19.1+6
- AA-Omniscience · Non-hallucinationGPT-5 by 5.917.811.9+5.9
Other results11 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 30.2-8.7-38.9+30.2
- AA Agentic IndexMiniMax M2.5 by 29.126.555.6+29.1
- swe_bench_bashMiniMax M2.5 by 10.86575.8+10.8
- τ²-Bench Telecom (AA run)MiniMax M2.5 by 10.584.895.3+10.5
- AIME25 no toolsGPT-5 by 8.394.686.3+8.3
- vectara_factual_consistencyMiniMax M2.5 by 684.990.9+6
- vectara_avg_summary_lengthGPT-5 by 25.5 rating points162.7137.2+25.5 rating
- ARC-AGI-1GPT-5 by 265.763.7+2
- vectara_answer_rateGPT-5 by 1.799.998.2+1.7
- Artificial Analysis Coding Indextie37.837.4tie
- AA Intelligencetie2322.8tie
Questions people ask
Which is better, GPT-5 or MiniMax M2.5?
GPT-5 wins four of the six areas where both have results: reasoning, facts, long documents and following instructions. MiniMax M2.5 wins coding and agents, and costs 87% less.
Which is better for coding?
MiniMax M2.5. It wins 3 of the 5 coding tests both models report; GPT-5 wins 1, and 1 is a tie.
Which is cheaper?
GPT-5 costs $1.25 per million input tokens and $10.00 per million output tokens; MiniMax M2.5 costs $0.30 and $1.20. That makes MiniMax M2.5 about 87% cheaper for the same work.
How do you compare the two?
We use the 27 benchmark tests both models have published scores on. The verdict counts the 16 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 11 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.