Kimi K3
Moonshot AIOpen weightsKimi K3 licenseReleased Jul 16, 2026
On the arena.ai Agents leaderboard, Kimi K3 has a task success rate of 1.95% (Max, rank #24, 95% interval 1.23%–2.66%). Kimi K3 ranks #18 on the LMArena Text leaderboard (style control) with an Arena score of 1487.1 (±4.6, 32,795 votes). Artificial Analysis Intelligence Index: 43.6. API price: $3.00 input / $15.00 output per 1M tokens. Context: 1.05M. Nearest rivals: Gemini 3.7 Flash (1487.6) above and Gemini 3.1 Pro Preview (1486.9) below.
Quick answer
- Agents leaderboard
- #24 of 48 · 1.95% success
- Text (LMArena)
- #18 · 1487.1
- 17NAS overall
- #24 of 84
- API price (in / out)
- $3.00 / $15.00 per 1M tokens
- Closest rival
- Gemini 3.7 Flash · 1487.6
- Cheaper alternative
- Gemini 3.8 Flash: Text 1496.4, output $3.75 (75% cheaper)
- Data updated
- Oct 16, 2026 · synced daily
Sources: arena.ai, LMArena, Artificial Analysis, OpenRouter
Key metrics
Reasoning effort comparison
Artificial Analysis measurements per reasoning effort of the same model| Effort | Intelligence | Coding | tok/s | TTFT (s) | First answer (s) |
|---|---|---|---|---|---|
| maxUsed in table | 43.6 | 76.2 | 45 | 2.8 | 47 |
| low | 30.1 | 72.0 | 43 | 2.9 | 50 |
Models ranked nearby
Sorted by Arena score| # | Model | Arena | Intel. | Output $ | tok/s |
|---|---|---|---|---|---|
| 12 |  Claude Opus 5 | 1490.0 | 50.8 | $25.00 | – |
| 13 |  Muse Spark | 1489.0 | 31.3 | – | – |
| 14 |  Gemini 3.7 Flash | 1487.6 | 39.1 | $3.75 | – |
| 15 | Kimi K3This page | 1487.1 | 43.6 | $15.00 | 45 |
| 16 | Gemini 3.1 Pro Preview | 1486.9 | 29.7 | $12.00 | 139 |
| 17 | Gemini 3 Pro | 1485.5 | 28.0 | $12.00 | – |
| 18 |  GPT-5.6 Sol | 1484.7 | 47.0 | $20.00 | – |
More from Moonshot AI
| # | Model | Arena | Intel. | Output $ | tok/s |
|---|---|---|---|---|---|
| 45 | Kimi K2.6 | 1460.8 | 27.0 | $4.00 | – |
| 55 | Kimi K2.5 | 1449.8 | 23.5 | $2.75 | – |
| – | Kimi K2.7 Code | – | 25.8 | $4.00 | 52 |
Pricing
- Official API list price: $3.00 input and $15.00 output per 1M tokens (USD).
- At a typical 3:1 input-to-output ratio, the blended cost is about $6.00 per 1M tokens.
- Among 79 models with an output price, it is the #64 cheapest.
- The cheapest OpenRouter route charges $0.30 input / $11.94 output. That route is over 60% below list price — usually a promotion or a single provider, so check its reliability.
Sources & methodology
On the arena.ai Agents leaderboard, Kimi K3 has a task success rate of 1.95% (Max, rank #24, 95% interval 1.23%–2.66%).
- arena.ai Agents — agent task success rate, 95% interval, sessions and rank (via the public Hugging Face dataset, CC BY 4.0).
- LMArena — Arena score (Text leaderboard with style control, the same view arena.ai shows by default), 95% confidence interval, votes and rank.
- Artificial Analysis — Intelligence / Coding Index, output speed and latency (medians).
- OpenRouter — Route prices, context length and listing date.
Arena scores use the Bradley-Terry model (not Elo); ± is the half-width of the 95% confidence interval. Latency figures are Artificial Analysis medians; for reasoning models, time to first token and to first answer include the thinking phase. Data syncs daily at 14:20 (UTC+8).
Limits: Arena scores come from anonymous users voting on chats, so they reflect conversational preference rather than agent or coding benchmarks; with few votes the interval is wide and ranks move. Prices and speeds change as providers update them.
FAQ
How does Kimi K3 do on the Agents leaderboard?
On the arena.ai Agents leaderboard, Kimi K3 has a task success rate of 1.95% (Max, rank #24, 95% interval 1.23%–2.66%).
What rank is Kimi K3 right now?
Kimi K3 is #15 on our leaderboard of 69 scored models (Arena score 1487.1); its raw LMArena rank is #18.
What does Kimi K3 cost via API?
Input $3.00 / output $15.00 per 1M tokens (official API list price; the OpenRouter quote when no official price is available). At a typical 3:1 input-to-output ratio, the blended cost is about $6.00 per 1M tokens.
How fast is Kimi K3?
Artificial Analysis measures about 45 output tokens per second, 2.8 s to the first token and 47 s to the first answer token; it is a reasoning model, so latency includes thinking time.
How close is Kimi K3 to the models ranked next to it?
It is 0.5 points behind Gemini 3.7 Flash (1487.6) and 0.2 points ahead of Gemini 3.1 Pro Preview (1486.9); gaps smaller than the confidence interval (±4.6) are not significant.
Is Kimi K3 open source? Can it be self-hosted?
Kimi K3 is an open-weights model (license: Kimi K3 license); you can download the weights and self-host.
What is the context length of Kimi K3?
1.05M (1,048,576 tokens, as listed on OpenRouter).
Kimi K3 vs Gemini 3.7 Flash: which is better?
On the LMArena Text leaderboard Kimi K3 scores 1487.1 and Gemini 3.7 Flash scores 1487.6 (a 0.5-point gap, within the confidence interval, so not significant). Agent task success: Kimi K3 1.95%, Gemini 3.7 Flash -0.41%. On price, Kimi K3 costs $15.00 and Gemini 3.7 Flash $3.75 per 1M output tokens.
Is there a cheaper model close to Kimi K3?
Yes: Gemini 3.8 Flash scores 1496.4 on the Text leaderboard (Kimi K3: 1487.1) and costs $3.75 per 1M output tokens, about 75% less than Kimi K3 ($15.00).
Where does the Kimi K3 data come from and how often is it updated?
Agent success rates come from the arena.ai Agents leaderboard (CC BY 4.0), Text scores from LMArena, intelligence and speed from Artificial Analysis, and prices and context from official list prices and OpenRouter. 17NAS syncs daily at 06:20 UTC; this page was last updated Oct 16, 2026.