Inference360
MiniMax

MiniMax M2.5

MiniMax M2.5 is a non-reasoning model from MiniMax. 10 benchmarks count toward its score, in 5 categories.

availableShows if the model has enough results for an index.
IndexOverall score. 50 is the middle.50.9 ±7.1
CoverageShare of the index weight with results.80%
SpeedOutput tokens per second.27/s
Input / 1MUS dollars per 1M input tokens.$0.27
Output / 1MUS dollars per 1M output tokens.$1.08
ContextMaximum tokens in one request.205K
EloLMArena rating and rank.1359 (#184)

50 is the middle of the board. The range shows the doubt in the index.

40,843 votes. Elo shows what people prefer. It does not change the score.

CapabilitiesScore per category. 50 is the middle.

50 is the middle
AgenticMulti-step tasks with tools.
52.3
CodingCode writing and repair.
51.5
ReasoningLogic problems and puzzles.
49.7
MultimodalTasks with images and text.
N/A
KnowledgeFacts and expert knowledge.
50.4
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
N/A
MathMath problems.
55.1

Results

10 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result on the index scale.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
AIMEMath88.8%55.1—16 Apr 2026Vals AI
GPQA DiamondKnowledge82.1%54.0—1 Sept 2026Vals AI
MMLU ProKnowledge80.1%46.7—1 Sept 2026Vals AI
LiveCodeBenchCoding79.2%56.3—1 Sept 2026Vals AI
SWE-bench VerifiedCoding75.8%58.3high effort · mini-SWE-agent1 Sept 2026SWE-bench team
SWE-benchCoding74.2%57.1—1 Sept 2026Vals AI
SWE-bench MultilingualMultilingual68.3%—mini-SWE-agent20 Feb 2026SWE-bench team
ARC-AGI-1 (semi-private)Reasoning63.7%57.3——ARC Prize Foundation
Terminal-Bench 2.0Agentic41.6%52.3—4 Jun 2026Vals AI
Vibe Code Bench v1.1Coding14.9%48.4OpenHands21 Sept 2026Vals AI
IOI v1Coding6.7%43.7—9 Aug 2026Vals AI
ARC-AGI-2 (semi-private)Reasoning4.9%42.2——ARC Prize Foundation

10 benchmarks count, from 11 of 12 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedVals AI, collected directlyNo licence stated. Read from the public leaderboard and credited to Vals AISWE-bench team, collected directlyNo licence stated. The repository publishes submission records for reproducibility and transparency and asks that SWE-bench be citedARC Prize Foundation, collected directlyNo licence stated. Their terms ask for written permission before commercial use

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeApodex 1.1 Mini59.1 · Free

More from MiniMax

MiniMax M356.9MiniMax M2.752.9MiniMax M2.150.0MiniMax 01UnrankedMiniMax 2.5Unranked