Compare · ModelsLive · 2 picked · head to head
MiniMax M2.5 vs MiniMax M3
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
MiniMax M3 wins on 10/12 benchmarks
MiniMax M3 wins 10 of 12 shared benchmarks. Leads in agentic · arena · coding.
Category leads
agentic·MiniMax M3arena·MiniMax M3coding·MiniMax M3reasoning·MiniMax M3language·MiniMax M3math·MiniMax M2.5knowledge·MiniMax M3general·MiniMax M3
Hype vs Reality
Attention vs performance
MiniMax M2.5
#131 by perf·#11 by attention
MiniMax M3
#124 by perf·#11 by attention
Best value
MiniMax M2.5
1.1x better value than MiniMax M3
MiniMax M2.5
72.0 pts/$
$0.68/M
MiniMax M3
66.5 pts/$
$0.75/M
Vendor risk
Mixed exposure
One or more vendors flagged
MiniMax
$4.0B·Tier 1
MiniMax
$4.0B·Tier 1
Head to head
12 benchmarks · 2 models
MiniMax M2.5MiniMax M3
APEX-Agents
MiniMax M3 leads by +31.5
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
MiniMax M2.5
6.2
MiniMax M3
37.7
Chatbot Arena Elo · Coding
MiniMax M3 leads by +95.8
MiniMax M2.5
1386.3
MiniMax M3
1482.0
Chatbot Arena Elo · Overall
MiniMax M3 leads by +49.2
MiniMax M2.5
1390.8
MiniMax M3
1440.1
LiveBench · Agentic Coding
MiniMax M3 leads by +8.3
MiniMax M2.5
51.7
MiniMax M3
60.0
LiveBench · Coding
MiniMax M2.5 leads by +2.5
MiniMax M2.5
70.7
MiniMax M3
68.2
LiveBench · Data Analysis
MiniMax M3 leads by +26.6
MiniMax M2.5
49.6
MiniMax M3
76.2
LiveBench · If
MiniMax M3 leads by +0.3
MiniMax M2.5
57.2
MiniMax M3
57.5
LiveBench · Language
MiniMax M3 leads by +21.7
MiniMax M2.5
55.1
MiniMax M3
76.8
LiveBench · Mathematics
MiniMax M2.5 leads by +0.5
MiniMax M2.5
77.4
MiniMax M3
77.0
LiveBench · Overall
MiniMax M3 leads by +9.9
MiniMax M2.5
60.1
MiniMax M3
70.0
LiveBench · Reasoning
MiniMax M3 leads by +15.2
MiniMax M2.5
59.3
MiniMax M3
74.5
Proofbench
MiniMax M3 leads by +14.0
MiniMax M2.5
4.0
MiniMax M3
18.0
Full benchmark table
| Benchmark | MiniMax M2.5 | MiniMax M3 |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 6.2 | 37.7 |
Chatbot Arena Elo · Coding | 1386.3 | 1482.0 |
Chatbot Arena Elo · Overall | 1390.8 | 1440.1 |
LiveBench · Agentic Coding | 51.7 | 60.0 |
LiveBench · Coding | 70.7 | 68.2 |
LiveBench · Data Analysis | 49.6 | 76.2 |
LiveBench · If | 57.2 | 57.5 |
LiveBench · Language | 55.1 | 76.8 |
LiveBench · Mathematics | 77.4 | 77.0 |
LiveBench · Overall | 60.1 | 70.0 |
LiveBench · Reasoning | 59.3 | 74.5 |
Proofbench | 4.0 | 18.0 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.27 | $1.08 | 205K tokens (~102 books) | $4.72 | |
| $0.30 | $1.20 | 1.0M tokens (~524 books) | $5.25 |