Skip to content
OSMEXO
Sign inGet an API key
☰
Sign inGet an API key

Kimi K3

moonshot/kimi-k3
  • Moonshot AI
  • Family · Kimi K3
  • Released 16 Jul 2026
  • Text · Image · Video → Text
  • Reasoning · adjustable
  • Tools · JSON · Vision
Live on OSMEXO
Try nowRequest a discount
API setup & price →
Spec & list price
Context1M
Max output944K
Input / 1M$0.19
Output / 1M$15.0078.1× input
Scores · standing
Intelligence43.6#25/489
LiveBench64.6#18/32
Coding76.2#10/196
Usage
Tokens / day201.6B▼ 18% 7d
Usage rank#24
Requests / day3.6M
Tool calls / req0.63
Per request · 30 days
Avg input59.3K4.0× median
Avg output1.0K1.5× median
Cache hit91%1.3× median
Reasoning share65%1.2× median
Tool-call failures0.3%0.2× median

Consumption

166.7B333.3B500.0B3 Oct17 Oct1 Nov
Peak 497.2B 13 OctLow 132.1B 1 NovRank today #24

News & trends

275 mentions this week ▲ 817%this model · videos, posts, articles · 30 days
Trending · 14 days
News

Benchmarks

bar = share of the top score · Benchmark data: Artificial Analysis
Long-context reasoning#1 of 393
88.7%
GPQA Diamond#10 of 445
93.5%
Coding index#10 of 196
76.2
Terminal-Bench 2.1#12 of 182
85.0%
τ-Bench Banking#9 of 154
46.0%
SciCode#13 of 160
59.5%
Humanity's Last Exam#22 of 465
46.9%
Intelligence index#25 of 489
43.6
Speed · median across 28 providers, latest day
Output speed#214/33453 tok/s1.3× slower than median
Fastest provider91 tok/sbest of 28 providers
First token#1941.28 s1.2× longer than median

#12 cheapest of 15 with a similar score

Value
Intelligence Index
4243444546$0.2$0.5$1$2$5Kimi K3
Blended price per 1M tokens (log)
LiveBench suites
Browser75.2
Coding59.3
Data Analysis78.7
Design77.2
Long Agentic26.4
Reasoning90.7
Research79.0

▲ Strongest: Reasoning, Research. ▼ Weakest: Long Agentic (26).

Technical details

Reasoning
on · max
Tokenizer
Other
License
other
Moderation
none
Downloads · likes
781.7K · 11.9K
Reasoning effort
  • max · default
  • high
  • low
  • none
Request parameters 22
  • tools
  • tool_choice
  • parallel_tool_calls
  • response_format
  • structured_outputs
  • reasoning
  • reasoning_effort
  • include_reasoning
  • temperature
  • top_p
+12 more
  • top_k
  • min_p
  • max_tokens
  • max_completion_tokens
  • stop
  • seed
  • frequency_penalty
  • logit_bias
  • logprobs
  • presence_penalty
  • repetition_penalty
  • top_logprobs

When to use

Best for

  • Hard reasoning and research questions
  • Coding agents with terminal workflows
  • Very long document analysis
  • Customer support with tool calling
  • Video and image understanding

Think twice for

  • Browser automation agents
  • Self-hosting (closed weights)
  • New projects (usage is declining)
  • High-volume simple chat (cheaper options exist)

Research papers

Same family

Kimi K3

Alternatives

other vendors · similar capability

Using Kimi K3 at volume?

Sign in to ask for a volume discount. You will continue in the OSMEXO dashboard.

Request a discount