Inference
EN

Claude Opus 5.5 vs MiniMax M3.1 Flash

Change models
Claude Opus 5.5Anthropic
MiniMax M3.1 FlashMiniMax
IndexAll categories, out of 100.
78.6 ±2.4 best
69.3 ±5.4
GapIndex points below the top.
Top
−9.3
CoverageIndex weight that has results.
95% best
80%
SpeedOutput tokens per second.
76/s
100/s best
EloLMArena rating and rank.
1513
—
ReleasedMonth of release.
2026-09
2026-10
Input / 1MUS dollars per 1M input tokens.
$4.00
$0.15 best
Output / 1MUS dollars per 1M output tokens.
$20.00
$0.60 best
Batch output / 1MBatch price per 1M output tokens.
$10.00
—
1K chats2K tokens in, 500 out, each.
$18.00
$0.60 best
1K documents20K tokens in, 1K out, each.
$100.00
$3.60 best
1K agent tasks200K tokens in, 20K out, each.
$1,200.00
$42.00 best
ContextMaximum tokens in one request.
1M
1.05M best
AgenticMulti-step tasks with tools.
74.1 best
71.8
CodingCode writing and repair.
81.8 best
69.4
ReasoningLogic problems and puzzles.
80.8 best
67.9
MultimodalTasks with images and text.
74.1 best
58.1
KnowledgeFacts and expert knowledge.
79.2 best
68.9
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
59.6
—
MathMath problems.
77.7
—

Benchmarks

28 shared
Only Claude Opus 5.5: 129Only MiniMax M3.1 Flash: 8
Claude Opus 5.5
MiniMax M3.1 Flash
Artificial Analysis Humanity's Last ExamKnowledge
61.4% best2 / 222
48.0%17 / 222
Artificial Analysis Long Context ReasoningReasoning
84.7% best6 / 217
79.7%45 / 217
Critical Physics TasksReasoning
31.7% best4 / 214
22.3%25 / 214
Artificial Analysis Omniscience AccuracyKnowledge
65.5% best3 / 164
45.1%37 / 164
GDPval-AA normalizedAgentic
68.3% best1 / 133
47.9%31 / 133
Vibe Code Bench v1.1Coding
90.3% best6 / 123
73.9%45 / 123
Artificial Analysis SciCodeCoding
66.9% best2 / 114
53.4%39 / 114
Artificial Analysis MMMU-ProMultimodal
87.7% best1 / 113
79.1%29 / 113
Legal Research BenchLegal
50.5% best7 / 90
34.1%50 / 90
Tax Agent BenchFinance
44.5% best6 / 82
32.1%23 / 82
LiveBench Data AnalysisReasoning
80.3% best9 / 80
71.4%52 / 80
LiveBench ReasoningReasoning
92.2% best4 / 80
82.9%48 / 80
Terminal-Bench 4.0Agentic
65.2% best1 / 61
25.3%28 / 61
DeepSWEAgentic
74.2% best10 / 59
73.7%13 / 59
Agent Arena steerabilityAgentic
18.6 best3 / 54
−2.636 / 54
Agent Arena task outcomeAgentic
20.7 best2 / 54
−0.734 / 54
Humanity's Last Exam without toolsKnowledge
64.4% best2 / 53
43.4%20 / 53
LiveBench Coding 2.0Coding
96.3% best3 / 52
83.9%26 / 52
LiveBench DesignAgentic
86.9% best1 / 52
74.0%31 / 52
LiveBench ResearchAgentic
85.0% best8 / 52
72.7%29 / 52