Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...
Context
1Mtokens
About 1,498 pages of text at once
Max output
944Ktokens
About 1,348 pages in one answer
Intelligence
17.4/ 100
Above 21% of the catalogue
Output speed
63tokens/s
A 500-word answer in about 11.2 s
From public market statistics for this model
First token
1.3s
Reasoning models start later
From public market statistics for this model
Health, 30 days
98.29%
17 of 30 days without a fault
IranRouter’s measurements
Price & cost
Every rate in Rial. A request pays only for the tokens it actually uses.
| Input | 526,271 | IRR / 1M tokens |
|---|---|---|
| Output | 41,114,895 | IRR / 1M tokens |
| Reasoning tokensThe output rate | 41,114,895 | IRR / 1M tokens |
| Cache read | 498,861 | IRR / 1M tokens |
| Image input tokensThe input rate | 526,271 | IRR / 1M tokens |
With prompt caching, the fixed part of your prompt is billed at the cache-read rate on later requests.
What would this model cost you a month?
Monthly cost
787.43MIRR
26.25M IRR a day, 13K IRR each request
The same work on peer models
30 days at today’s rate.
Rial price over 90 days
It moves with the exchange rate; the steps are the maker’s own price changes. Built at today’s margin, so approximate if the margin changed in the window.
Performance, 30 days
Health is measured by IranRouter itself: a day counts as healthy unless requests to this model failed on our side or it could not be served. First token and speed come from IranRouter’s traffic, and from public market statistics for this model on days without enough of it.
Daily health98.29%
Time to first token1.47 s
Output speed55 tokens/s
Place in the market
By usage across the global model market.
Daily rank#25
Share of the day’s tokens0.8%
Among its peers
Cost of one message against intelligence; this model is highlighted.
Cost of one message, IRR (log scale)
Capabilities
- Input
- TextImageVideo
- Output
- Text
Parameters it accepts
frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_completion_tokensmax_tokensmin_pparallel_tool_callspresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_pOther benchmarks
Design contests
- 3DElo: 1,382Win rate: 64%
- Code (overall)Elo: 1,352Win rate: 60%
- Data vizElo: 1,351Win rate: 63%
- Game devElo: 1,348Win rate: 54%
3 moreShow less
- UI componentsElo: 1,338Win rate: 60%
- WebsitesElo: 1,334Win rate: 59%
- SVGElo: 1,305Win rate: 60%
Standard tests
- GPQA DiamondScore: 92%Tasks: 2,175
- τ-bench Verified AirlineScore: 73%Tasks: 500
Start in a minute
The official OpenAI and Anthropic SDKs; only the address and the key change.
# pip install openai from openai import OpenAI client = OpenAI( base_url="https://iranrouter.com/v1", api_key="ir-...", ) r = client.chat.completions.create( model="moonshotai/kimi-k3", messages=[{"role": "user", "content": "Hello, introduce yourself"}], ) print(r.choices[0].message.content)
Peer models
Its family and models with a close score; the last column compares one message’s cost with this model.
| Model | Against this model | ||||
|---|---|---|---|---|---|
| MiniMax: MiniMax M3Minimax | 3,289 | 18.7 | 87 | 1M | 85% cheaper |
| Qwen: Qwen3.8 27BQwen | 3,371 | 20.1 | 51 | 1M | 84% cheaper |
| NVIDIA: Nemotron 3 UltraNVIDIA | 6,578 | 11.8 | 69 | 262K | 70% cheaper |
| Google: Gemini 3.5 Flash LiteGoogle | 5,071 | 11.7 | 132 | 1M | 77% cheaper |
| DeepSeek: DeepSeek V4 Pro 0813DeepSeek | 11,531 | 23.1 | 52 | 1M | 47% cheaper |
Technical specs
- ID
moonshotai/kimi-k3- Maker
- Moonshot AI
- Tokenizer
- Other
- Released
- Jul 16, 2026
- Added to IranRouter
- Aug 19, 2026
- Context
- 1,048,576 tokens
- Max output
- 943,718 tokens
- Weights
- moonshotai/Kimi-K3
Default parameters
top_p=0.95Scores by Artificial Analysis, Design Arena, as of Oct 31, 2026