Kimi K3.1 Pro
Oct 2026
Kimi K3.1 ProCompare inon
Benchmarks of Kimi K3.1 Pro8
BenchmarkByScoreRank
Math5
Writing1
Health & law1
Legacy · saturated or no longer reported1
No benchmark matches
Rankingat high effort?
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperGrok 4.7xhighwins 46%-25%$3.24Context: 1M → 500K+5
BetterGPT-6.1 Solmaxwins 53%same$0.72Remove unsupported parameters+5OpenKimi K3unstatedsame+50%$3 / $15Sampling parameters are fixed+3
CheaperMuse Spark OpenunstatedOpenwins 45%-50%$1.25 / $4.25Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1
CheaperGemini 3.8 Flashunstatedwins 45%-63%*$0.75 / $3.75Sampling parameters are ignored+1ModelScoreCostPer taskWhat changes
Grok 4.7xhighwins 46%-25%$3.24Context: 1M → 500K+5Muse Spark OpenunstatedOpenwins 45%-50%$1.25 / $4.25Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1Gemini 3.8 Flashunstatedwins 45%-63%*$0.75 / $3.75Sampling parameters are ignored+1Muse Spark 1.3unstatedwins 44%-50%$1.25 / $4.25Only auto tool_choice is supported+5
Mistral Large 4unstatedwins 42%-48%$1.36 / $4.18Infinite whitespace stream+4ModelScoreCostPer taskWhat changes
GPT-6.1 Solmaxwins 53%same$0.72Remove unsupported parameters+5GPT-6 Astramaxwins 53%+400%$3.26Monitoring can return HTTP 403+2
Claude Fable 5.1highwins 52%+400%$3.91Replace forced tool_choice+5ModelScoreCostPer taskWhat changes
Kimi K3unstatedOpensame+50%$3 / $15Sampling parameters are fixed+3Muse Spark OpenunstatedOpenwins 45%-50%$1.25 / $4.25Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1ModelScoreCostPer taskWhat changes
Kimi K3unstatedOpensame—$3 / $15Sampling parameters are fixed+3Kimi K3.1 Pro Highspeedn/a—$4 / $20Kimi K2.7 CodethinkingOpenn/a—$0.54Context: 1M → 256K+1Kimi K2.6thinkingOpenn/a—$0.71Context: 1M → 256KKimi K2.5thinkingOpenn/a—$0.60 / $3Context: 1M → 256KKimi K2 ThinkingthinkingOpenn/a—$0.47 / $2Context: 1M → 256K+1Kimi K2 0905non-thinkingOpenn/a—$0.40 / $2Context: 1M → 256K+1Kimi K2 0711non-thinkingOpenn/a—$0.50 / $2Context: 1M → 128K+2All Moonshot AI models →Benchmarks
all categoriesCoding
Terminal-Bench 4.0lab65.7
Terminal-Bench Science 0.1lab38.1
SWE Marathonlab51.7
Kimi Code Bench 3.0 Long-horizonlab61.2
Agents & Tool Use
GDPVal-AA v2.1lab1,846
OSWorld V2.1lab79.6
Agents Last Examlab43.2
OfficeQA Pro v2lab64.4
Frontier Financelab58.6
Writing & Language
Vision & Documents
BenchCAD w/ toolslab92.7
FAQ
How much does Kimi K3.1 Pro cost?›
$2 per 1M input tokens and $10 per 1M output tokens; cached input $0.20 per 1M.
Is Kimi K3.1 Pro free to use?›
There is no free API; the cheapest paid price is $2 per 1M input tokens.
How good is Kimi K3.1 Pro, and what is it best at?›
#36 of 73 in Chat & Knowledge, #7 of 98 in Coding.
Is Kimi K3.1 Pro better than Claude Opus 5.5?›
On MathArena · BrokenArXiv · August 2026, Kimi K3.1 Pro scores 64.3 vs 77.4 — about the same.
How can I use Kimi K3.1 Pro?›
Through the API of Moonshot AI, Novita.
What is the context window of Kimi K3.1 Pro?›
1M tokens.
When was Kimi K3.1 Pro released? Is there a newer version?›
Released 2026-10-23. It is the current version in its line.
Is Kimi K3.1 Pro open source?›
No, Kimi K3.1 Pro is a closed model available through an API. The closest open-weights option here is Kimi K3.
Can Kimi K3.1 Pro read images?›
It accepts text and images, video as input.
What is Kimi K3.1 Pro?›
Kimi K3.1 Pro is a closed language model from Moonshot AI, released Oct 2026.
What are the best alternatives to Kimi K3.1 Pro?›
Grok 4.7 (cheaper), GPT-6.1 Sol (better), Kimi K3 (open weights).