Inference
EN

Gemini 4 Carbon vs MiMo-V2.6-Pro

Change models
Gemini 4 CarbonGoogle
MiMo-V2.6-ProXiaomi
IndexAll categories, out of 100.
82.7 ±5.1 best
70.2 ±4.7
GapIndex points below the top.
Top
−12.5
CoverageIndex weight that has results.
80%
80%
SpeedOutput tokens per second.
31/s
31/s
EloLMArena rating and rank.
1538 best
1492
ReleasedMonth of release.
2026-10
2026-09
Input / 1MUS dollars per 1M input tokens.
$2.00
$0.435 best
Output / 1MUS dollars per 1M output tokens.
$10.00
$0.87 best
Batch output / 1MBatch price per 1M output tokens.
$5.00
—
1K chats2K tokens in, 500 out, each.
$9.00
$1.31 best
1K documents20K tokens in, 1K out, each.
$50.00
$9.57 best
1K agent tasks200K tokens in, 20K out, each.
$600.00
$104.40 best
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
83.2 best
72.1
CodingCode writing and repair.
80.5 best
66.6
ReasoningLogic problems and puzzles.
83.9 best
68.3
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
79.5 best
64.2
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
77.2 best
71.3

Benchmarks

15 shared
Only Gemini 4 Carbon: 24Only MiMo-V2.6-Pro: 43
Gemini 4 Carbon
MiMo-V2.6-Pro
Artificial Analysis Long Context ReasoningReasoning
79.3%52 / 217
86.3% best3 / 217
Artificial Analysis Omniscience AccuracyKnowledge
58.3% best9 / 164
34.8%67 / 164
Vibe Code Bench v1.1Coding
89.1% best9 / 123
85.2%16 / 123
Legal Research BenchLegal
55.8% best1 / 90
47.1%16 / 90
Tax Agent BenchFinance
51.8% best1 / 82
32.4%22 / 82
Terminal-Bench 4.0Agentic
63.1% best4 / 61
31.3%24 / 61
Agent Arena steerabilityAgentic
19.1 best2 / 54
1.525 / 54
Agent Arena task outcomeAgentic
18.4 best4 / 54
5.016 / 54
HLAB-Vals-1.1Legal
17.5% best8 / 35
13.8%13 / 35
AutomationBenchAgentic
53.9% best5 / 32
53.1%7 / 32
Artificial Analysis AutomationBenchAgentic
77.8% best1 / 23
58.6%15 / 23
ExploitGymAgentic
53.5% best1 / 23
17.8%14 / 23
Artificial Analysis GDP.pdfAgentic
28.4% best7 / 22
19.2%18 / 22
Artificial Analysis Terminal-Bench v4.0Agentic
65.2% best2 / 22
34.8%14 / 22
Multimodal Knowledge Agent BenchmarkAgentic
54.3% best2 / 19
37.3%13 / 19