BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases12
Moonshot AIReasoning modelNewReleased 23 Oct 2026Compare with another model

Kimi K3.1 Pro

Kimi K3.1 Pro is a Moonshot AI proprietary reasoning model, released 23 Oct 2026. Its best configuration ranks #29 of 785 on the BenchLeader Index at 66.8 ±6.2, in the upper half. The ± is the point: 111 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (81) and lowest in multimodal (67). At $4.00 per million tokens blended it is among the most expensive ranked models. Output speed of 81 tokens per second puts it slower than most, with a first answer in 27.7 s. Last measured 23 Oct 2026.

Blended price
$4.00/M
$2.00 in · $10.00 out
Output speed
81 tok/s
measured by Artificial Analysis
First answer
28 s
first token 4.06 s
Context
1.0M
Full answer
34 s
median, reasoning included
Cost per run
$2.11
one full Intelligence Index run
Released
23 Oct 2026
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index67
  2. Reasoning81
  3. Agents & tools67
  4. Knowledge75
  5. Multimodal67
  6. Long context68
  7. Composite76

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmarknot statedSource
LiveBench 2.0 Reasoningnot in index91.1%#10LiveBench
Humanity's Last Exam (AA)not in index52.5%#22Artificial Analysis
CritPt24.6%#47Artificial Analysis

Coding

Benchmarknot statedSource
LiveBench 2.0 Codingnot in index81.8%#29LiveBench
SciCode (AA)not in index60.8%#10Artificial Analysis
Terminal-Bench 4.0 (AA)not in index67.2%#1Artificial Analysis
FrontierCode 1.1 (AA)not in index47.1%#7Artificial Analysis

Agents & tools

Benchmarknot statedSource
LiveBench Long Agenticnot in index49.5%#14LiveBench
LiveBench Browsernot in index80.0%#16LiveBench
GDPval-AA v2.1not in index67.3%#3Artificial Analysis
AutomationBenchnot in index73.3%#2Artificial Analysis
GDP.pdf (AA)not in index31.4%#5Artificial Analysis
AA-Briefcase v1.1not in index1678#7Artificial Analysis
CWE-Bench (AA)60.8%#7Artificial Analysis
DeepSec-Bench (AA)36.7%#3Artificial Analysis
CyberGym end-to-end (AA)94.7%#1Artificial Analysis

Knowledge

Benchmarknot statedSource
LiveBench 2.0 Data Analysisnot in index79.1%#16LiveBench
LiveBench Researchnot in index77.9%#22LiveBench
AA-Omniscience29.5#35Artificial Analysis
AA-Omniscience: accuracynot in index51.0%#60Artificial Analysis
AA-Omniscience: non-hallucinationnot in index56.1%#90Artificial Analysis

Multimodal

Benchmarknot statedSource
LiveBench Designnot in index77.8%#21LiveBench
MMMU-Pro83.9%#24Artificial Analysis

Long context

Benchmarknot statedSource
AA-LCR86.3%#3Artificial Analysis

Composite

Benchmarknot statedSource
LiveBench 2.076.7%#14LiveBench
AA Intelligence Index v4.3.255.1#5Artificial Analysis
AA Cyber Indexnot in index64.1#1Artificial Analysis

Where it wins

Benchmarks where this configuration ranks in the top five of every configuration measured.

Where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

ProviderSpeedFirst tokenInput $/MOutput $/MContextQuantisation
Moonshot AI52 tok/s3.36 s$4.00$20.001.0Mmxfp4
Moonshot AI43 tok/s4.06 s$2.00$10.001.0Mmxfp4

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.200 per 1M. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.0038$0.003331.5 s
Summarize a 30-page report12,000 / 600$0.030$0.01435.2 s
Code edit6,000 / 1,500$0.027$0.01946.3 s
Agentic coding session60,000 / 4,000$0.160$0.0791.3 min
Structured extraction2,000 / 200$0.0060$0.003330.2 s

More from Moonshot AI

New this month

Compare Kimi K3.1 Pro with…

See also

Best LLM for agentsCheapest 1M contextBest for agentsAll Moonshot AI models

Questions people ask

What are Kimi K3.1 Pro's benchmark scores?
As of 24 Oct 2026, Kimi K3.1 Pro scores 66.8 ±6.2 on the BenchLeader Index, top 50, strongest in reasoning (81) and weakest in multimodal (67). That rests on 9 independent benchmark results, and every score on this page links to its publisher.
How does Kimi K3.1 Pro rank?
Kimi K3.1 Pro is #29 of 785 ranked configurations on 24 Oct 2026 (top 50). 111 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
How much does Kimi K3.1 Pro cost per 1M tokens?
As of 24 Oct 2026, Kimi K3.1 Pro costs $2.00 per 1M input tokens and $10.00 per 1M output tokens, $4.00 blended (three input tokens per output token).
Is Kimi K3.1 Pro better than Muse Spark 1.3?
As of 24 Oct 2026, Kimi K3.1 Pro and Muse Spark 1.3 (max) are neck and neck on quality (66.8 vs 67.1 on the BenchLeader Index, inside the ±7.6 margin of error), and Muse Spark 1.3 (max) costs 2.0× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.
When was Kimi K3.1 Pro released?
Moonshot AI released Kimi K3.1 Pro on 23 Oct 2026.

Data as of 24 Oct 2026. Compare with another model.

Cite as: BenchLeader, “Kimi K3.1 Pro: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/kimi-k3-1-pro.