Kimi K3

Moonshot AIKey kimi-k3Also reported as moonshotai/kimi-k3, alibaba/kimi-k3, combo/kimi-k3, kimi/kimi-k3, cline-pass/kimi-k3, merius/moonshotai/kimi-k3, hf:moonshotai/kimi-k3, modal/kimi-k3, inferhub/kimi-k3, aihubmix/kimi-k3, fact/kimi-k3, sk/kimi-k3, cline-free/kimi-k3, moonshot/kimi-k3, cb/kimi-k3, newapi/kimi-k3, ali/kimi-k3, synthetic/hf:moonshotai/kimi-k3, openrouter/kimi-k3, cline/kimi-k3

Rank
#51
by active users
Active users
7
Down 10 from 17 active users in the previous 7 days. vs previous 7 days
Adoption
1.2%
of 598 accounts
Tokens
36.3M
cache included
Est. cost
$27
at public list prices
First seen
Jun 13, 2026
in any submission

Nov 7–13, 2026 (UTC)598 accounts100% of tokens from an identified modelData through Nov 13, 2026, 01:19 UTCHow we count

Daily active users

Distinct accounts that used this model on each UTC day.

1 active users on Nov 13, 2026

Daily active users for Kimi K3, Nov 7, 2026 to Nov 13, 2026. Peak 3. Use the left and right arrow keys to read each day.320
Show daily values
Daily active users and tokens for Kimi K3
Date (UTC)Active usersTokens
Nov 7, 20262978K
Nov 8, 202629.7M
Nov 9, 2026318.1M
Nov 10, 202624.6M
Nov 11, 20261238K
Nov 12, 202622.3M
Nov 13, 20261426K

Token mix

How the 36.3M tokens split. Cache reads are usually most of an agent's tokens and are priced far below input.

  • Input9.7%
  • Output1.4%
  • Cache read88%
  • Cache write0.7%
  • Reasoning<0.1%

Agents that carry Kimi K3

Each agent's share of this model's tokens, and how many accounts used the model through it.

  1. Claude Code63% of this model's tokens3active users
  2. Senpi (OmO Native)27% of this model's tokens1active users
  3. Amp5.8% of this model's tokens1active users
  4. Oh My Pi4.9% of this model's tokens1active users
  5. cc-mirror/curl<0.1% of this model's tokens1active users
  6. cc-mirror/python-urllib<0.1% of this model's tokens1active users
  7. cc-mirror/mozilla<0.1% of this model's tokens1active users

Benchmarks

The best result for Kimi K3 in each benchmark Tokscale tracks. Each benchmark has its own unit, run conditions and source, so compare a score only with other models on the same benchmark. Rank is among the models Tokscale matched in that benchmark.

All benchmarks
  • Epoch Capabilities IndexOverall · Epoch AI
    158.6
    Rank #24 of 158 matched modelsNo conditions reportedModel released Jul 16, 2026Source: epoch.ai (opens in a new tab)
  • DeepSWEAgentic coding · Datacurve, collected by Epoch AI
    68.5%
    Rank #8 of 26 matched modelsMax · mini-swe-agentModel released Jul 16, 2026Source: deepswe.datacurve.ai (opens in a new tab)
  • FrontierSWEAgentic coding · FrontierSWE, collected by Epoch AI
    25.9%
    Rank #14 of 23 matched modelsMax · proximusModel released Jul 16, 2026Source: frontierswe.com (opens in a new tab)
  • WebDev ArenaWeb development · LMArena
    1,655
    Rank #14 of 110 matched modelsMaxPublished Oct 19, 2026Source: huggingface.co (opens in a new tab)
  • Agent ArenaAgents · LMArena
    +0.02
    Rank #23 of 48 matched modelsMax · Arena Agent ModePublished Oct 19, 2026Source: huggingface.co (opens in a new tab)
  • GPQA DiamondReasoning · Epoch AI
    93.1%
    Rank #24 of 160 matched modelsMaxRun date Jul 16, 2026Source: epoch.ai (opens in a new tab)

    Near its ceiling: the top score is 96.5%, and Epoch AI estimates the reachable maximum at 97%. Differences between top models are small.

    Show 2 more results with other run conditionsHide the other results
    • Listed as kimi-k3_high
      91.9%
      HighRun date Aug 7, 2026Source: epoch.ai (opens in a new tab)
    • Listed as kimi-k3_low
      84.8%
      LowRun date Aug 7, 2026Source: epoch.ai (opens in a new tab)
  • FrontierMath Tiers 1–3 (v2)Reasoning · Epoch AI
    72.2%
    Rank #28 of 90 matched modelsMaxRun date Jul 17, 2026Source: epoch.ai (opens in a new tab)

    Near its ceiling: the top score is 95.1%, and Epoch AI estimates the reachable maximum at 95%. Differences between top models are small.

  • SimpleQA VerifiedKnowledge · Epoch AI
    50.6%
    Rank #30 of 86 matched modelsMaxRun date Aug 27, 2026Source: epoch.ai (opens in a new tab)

Not evaluated: Humanity's Last Exam, Terminal-Bench 4.1, CursorBench. A missing result is not a score of zero.

Pricing

List prices from models.dev in US dollars per 1M tokens, retrieved Nov 12, 2026. Moonshot AI is the lab's own API and the reference price for the estimates below.

Price at real agent usageEstimate at list prices

What 1M tokens cost at the reference list price when they split into input, output and cache the way Tokscale users' agents used them. It is not what anyone paid.

At this model's mix
$0.78
Est. per 1M tokens
88% cache reads
At the community mix
$0.45
Est. per 1M tokens
96% cache reads
3:1 input/output blend
$6
Per 1M tokens at list prices
No cache

Cache reads were 88% of Kimi K3 tokens in the last 7 days and 96% of all tokens. The 3:1 blend, (3 × input + output) ÷ 4, is the usual comparison and ignores cache.

Assumptions

  • Moonshot AI list prices per 1M tokens: input $3, output $15, cache read $0.30, cache write $3.
  • This model's mix is its own input, output, cache and reasoning tokens in the last 7 days.
  • The community mix is the tokens of every identified model in the last 7 days.
  • This is an estimate at list prices, not what anyone paid.
  • Reasoning tokens are priced as output.
  • Long-context tiers, batch pricing and discounts are not applied.
  • Token mixes come from public Tokscale submissions. Hidden accounts are excluded.

List prices by provider

  • Moonshot AIReferencekimi-k3
    $3 input per 1M tokens$15 output per 1M tokens$0.30 cache read per 1M tokens$3 cache write per 1M tokens1.05M tokens of context
  • Amazon Bedrockus.moonshotai.kimi-k3
    $3.30 input per 1M tokens$16.50 output per 1M tokens$0.33 cache read per 1M tokens$4.13 cache write per 1M tokens1.05M tokens of context
  • OpenRoutermoonshotai/kimi-k3
    $0.19 input per 1M tokens$16 output per 1M tokens$0.18 cache read per 1M tokensCache write not listed1.05M tokens of context
  • Vercel AI Gatewaymoonshotai/kimi-k3
    $3 input per 1M tokens$15 output per 1M tokens$0.30 cache read per 1M tokensCache write not listed1M tokens of context
Show 4 more providersShow fewer providers
  • Together AImoonshotai/Kimi-K3
    $3 input per 1M tokens$15 output per 1M tokens$0.30 cache read per 1M tokensCache write not listed1.05M tokens of context
  • Fireworks AIaccounts/fireworks/models/kimi-k3
    $3 input per 1M tokens$15 output per 1M tokens$0.30 cache read per 1M tokensCache write not listed1.05M tokens of context
  • Deep Inframoonshotai/Kimi-K3
    $2.85 input per 1M tokens$14.25 output per 1M tokens$0.28 cache read per 1M tokens$3.56 cache write per 1M tokens1.05M tokens of context
  • Basetenmoonshotai/Kimi-K3
    $3 input per 1M tokens$15 output per 1M tokens$0.30 cache read per 1M tokensCache write not listed1.05M tokens of context

Specs

From models.dev (MIT), retrieved Nov 12, 2026, as the lab or the first listed provider reports them. A dash means models.dev does not list the value.

Context window
1.05M tokens
Max output
1.05M tokens
Input
Text, Image, Video
Output
Text
Knowledge cutoff
— not listed
Release date
Jul 16, 2026
Open weights
Yes
Reasoning
Yes
Tool calling
Yes

Top public profiles

Public profiles that used Kimi K3 most in this period, by tokens. Hidden accounts are never listed.

  1. 1Christopher Butler@techsaint17.3Mtokens$11est. cost3active days$11 est. cost · 3 active days
  2. 2DevNewbie1826@DevNewbie18269.7Mtokens$7est. cost1active days$7 est. cost · 1 active days
  3. 3Park Hee Chan@channprj5.3Mtokens$5est. cost1active days$5 est. cost · 1 active days
  4. 4Rakai Seto@RakaiSeto2.1Mtokens—est. cost2active daysno price · 2 active days
  5. 5崔鹏飞 Cui Pengfei@cuipengfei1.8Mtokens$3est. cost3active days$3 est. cost · 3 active days
  6. 6曹亚文@bobcyw79.9Ktokens—est. cost1active daysno price · 1 active days
  7. 7如是@MiFaZhan3.4Ktokens<$1est. cost2active days<$1 est. cost · 2 active days

More from Moonshot AI

Other Moonshot AI models in the same period, by active users.

  1. Kimi K3.1 Pro2.8% adoption17active users
  2. Kimi For Coding0.8% adoption5active users
  3. Kimi For Coding Highspeed0.3% adoption2active users
  4. Kimi K2.60.3% adoption2active users
  5. Kimi K3 Max0.3% adoption2active users
  6. Kimi K2.7 Code0.3% adoption2active users
All models

Sources and attribution

Each score keeps its own source, unit and run conditions. Tokscale does not combine benchmarks into one score, and a missing score means the model was not evaluated, not that it scored zero.

  • Epoch AIEpoch Capabilities Index, Epoch's own runs and the external results it collects.CC BY 4.0 · Retrieved Nov 12, 2026 · Data through Nov 9, 2026Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from https://epoch.ai/benchmarks.
  • LMArenaWebDev Arena and Agent Arena leaderboards.CC BY 4.0 · Retrieved Oct 21, 2026 · Data through Oct 19, 2026LMArena leaderboard dataset (lmarena-ai/leaderboard-dataset), Hugging Face.
  • models.devSpecs and list prices per provider.MIT · Retrieved Nov 12, 2026models.dev, an open catalog of AI models (github.com/anomalyco/models.dev).
  • TokscaleActive users, tokens and token mix from public Tokscale submissions. Hidden accounts are excluded.How we count