Inference
EN

GPT-5.6 Sol vs Kimi K3.1 Pro

Change models
GPT-5.6 SolOpenAI
Kimi K3.1 ProMoonshot AISmartest open
IndexAll categories, out of 100.
73.0 ±1.9
75.7 ±4.7 best
GapIndex points below the top.
−2.7 · no clear gap
Top
CoverageIndex weight that has results.
100% best
80%
SpeedOutput tokens per second.
64/s best
43/s
EloLMArena rating and rank.
1457
—
ReleasedMonth of release.
2026-07
2026-10
Input / 1MUS dollars per 1M input tokens.
$4.40
$2.00 best
Output / 1MUS dollars per 1M output tokens.
$22.00
$10.00 best
Batch output / 1MBatch price per 1M output tokens.
$10.00
$6.00 best
1K chats2K tokens in, 500 out, each.
$19.80
$9.00 best
1K documents20K tokens in, 1K out, each.
$110.00
$50.00 best
1K agent tasks200K tokens in, 20K out, each.
$1,320.00
$600.00 best
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
71.9
76.7 best
CodingCode writing and repair.
70.4
71.9 best
ReasoningLogic problems and puzzles.
76.3
77.7 best
MultimodalTasks with images and text.
72.6 best
72.0
KnowledgeFacts and expert knowledge.
72.7 best
71.0
MultilingualTasks in many languages.
66.4
—
InstructionTasks with strict prompt rules.
64.5
—
MathMath problems.
75.1
—

Benchmarks

19 shared
Only GPT-5.6 Sol: 106Only Kimi K3.1 Pro: 44
GPT-5.6 Sol
Kimi K3.1 Pro
Artificial Analysis Humanity's Last ExamKnowledge
49.5%13 / 222
52.5% best11 / 222
Artificial Analysis Long Context ReasoningReasoning
84.0%8 / 217
86.3% best3 / 217
Critical Physics TasksReasoning
32.3% best3 / 214
24.6%22 / 214
Artificial Analysis Omniscience AccuracyKnowledge
59.4% best8 / 164
48.9%23 / 164
Graduate-Level Google-Proof Q&AKnowledge
94.6% best4 / 141
94.2%5 / 141
GDPval-AA normalizedAgentic
55.4%19 / 133
67.3% best3 / 133
Artificial Analysis SciCodeCoding
57.1%22 / 114
60.8% best9 / 114
Artificial Analysis MMMU-ProMultimodal
83.4%10 / 113
83.9% best9 / 113
GPQA Diamond · BenchLMKnowledge
94.6% best4 / 84
94.2%6 / 84
BrowseCompAgentic
92.2% best2 / 63
92.1%3 / 63
DeepSWEAgentic
72.7%15 / 59
75.5% best5 / 59
Massive Multi-discipline Multimodal Understanding ProMultimodal
83.0%5 / 53
84.1% best2 / 53
MazeBench: visual spatial reasoning in a 3D open worldAgentic
13.0% best5 / 48
2.0%15 / 48
FrontierCode 2.0Coding
27.6% best12 / 27
19.6%22 / 27
Artificial Analysis AnalystAgentAgentic
47.5% best14 / 26
38.8%23 / 26
Bug Hunt BenchCoding
42.0% best6 / 25
36.0%11 / 25
OSWorld 2.0Agentic
62.6%6 / 25
79.6% best1 / 25
ExploitGymAgentic
33.7% best7 / 23
20.7%13 / 23
MMMU-Pro with PythonMultimodal
84.6%
90.0%