kimi-k3
Moonshot AI · text, image, video → text
Kimi K3 is Kimi’s flagship model for long-horizon coding and end-to-end knowledge work, with a 1M-token context window and industry-leading intelligence.
Input$3.00 /M
Output$15.00 /M
← Modelskimi-k3
qwen3.8-max
kimi-k3 vs qwen3.8-max
Output tokens cost $15.00 per million on kimi-k3 and $5.07 per million on qwen3.8-max. Input tokens cost $3.00 per million on kimi-k3 and $1.69 per million on qwen3.8-max. Cached input tokens are billed at $0.30 per million on kimi-k3 and $0.17 per million on qwen3.8-max. Context windows are 1,048,576 tokens on kimi-k3 and 991,000 tokens on qwen3.8-max. First-token latency measured on AIHubMix is 6.8s on kimi-k3 and 3.5s on qwen3.8-max. Measured output throughput is 32.2 tokens/s on kimi-k3 and 37.1 tokens/s on qwen3.8-max. On the LMArena coding leaderboard kimi-k3 scores 1531 and qwen3.8-max scores 1530.
kimi-k3qwen3.8-max+ Add model 2/4
qwen3.8-max
Qwen · text → text
Qwen 3.8 Max(qwen3.8-max) is Alibaba Cloud’s flagship native vision-language model, built on a 2.4-trillion-parameter Mixture-of-Experts (MoE) architecture and supporting context windows of up to 1 million tokens. It is well suited for complex multimodal understanding, advanced reasoning, software development, agentic workflows, and long-context processing. At a similar price to Qwen3.7-Max, Qwen3.8-Max delivers significant improvements in reasoning, coding, and agent capabilities, with overall performance comparable to today’s leading models.
Input$1.69 /M
Output$5.07 /M
Specs & Pricing
Prices are per million tokens. Latency and throughput are rolling averages measured on AIHubMix.
kimi-k3
qwen3.8-max
Input price
$3.00
$1.69
Output price
$15.00
$5.07
Cache read
$0.30
$0.17
Context window
1,048,576
991,000
Max output
1,048,576
128,000
Latency (first token)
6.8 s
3.5 s
Throughput
32.2 TPS
37.1 TPS
Modalities
textimagevideo
text
Features
thinkingfunction callingstructured outputs
toolsfunction callingstructured outputsweblong contextthinking
Endpoints
Promotional prices show the discounted rate; see each model page for promotion windows.
Usage on AIHubMix last 30 days
Daily traffic served through AIHubMix — how demand for each model is trending.
kimi-k3qwen3.8-max
Tokens / day
-
Requests / day
-
Live Performance last 3 days
Measured on real AIHubMix traffic, hourly buckets. Gaps mean no traffic in that hour.
kimi-k3qwen3.8-max
Throughput (TPS)
-
Latency (s)
-
Uptime (%)
-
Arena Scores
LMArena ratings by capability (Bradley-Terry, commonly called Elo). Higher is better.
Text
kimi-k3qwen3.8-max
1440148015201560
Overall
14861496
Coding
15301531
Hard prompts
15051518
Instruction following
14791486
Multi-turn
14991507
Creative writing
14701493
Longer query
15021510
Chinese
15271560
English
14911503
Source: LMArena (arena.ai) leaderboard, imported by AIHubMix. Models without published ratings are omitted per chart.
Cost Calculator
Estimate your monthly bill for the same workload on each model.
qwen3.8-max
$177 /mo
kimi-k3
$405 /mo
Monthly = daily × 30. Discounted rates applied where a promotion is active.
FAQ
Which is cheaper: kimi-k3, qwen3.8-max?
qwen3.8-max: $5.07/M output tokens; kimi-k3: $15.00/M. Use the cost calculator above to estimate your own workload.
How do their coding arena scores compare?
kimi-k3: 1531; qwen3.8-max: 1530 (LMArena coding leaderboard).
Which responds faster?
qwen3.8-max: 3.5s first-token latency measured on AIHubMix; see the live performance charts above for how each model behaves across the day.
How large is each context window?
kimi-k3 accepts 1,048,576 and qwen3.8-max accepts 991,000 input tokens. Maximum output per request is 1,048,576 tokens on kimi-k3 and 128,000 tokens on qwen3.8-max.
Which one generates tokens faster?
qwen3.8-max at 37.1 tokens/s and kimi-k3 at 32.2 tokens/s, measured as output throughput on AIHubMix — a separate metric from first-token latency.
What inputs and capabilities does each model support?
kimi-k3 accepts text, image and video input and supports thinking, function calling and structured outputs; qwen3.8-max accepts text input and supports tool calling, function calling, structured outputs, web search, long context and thinking.
Can I call kimi-k3 and qwen3.8-max with the same API key?
Yes. AIHubMix serves every model on this page behind one OpenAI-compatible endpoint, so switching between them is a one-line change to the model field — no second account, key or SDK.
Popular Comparisons
Related model match-ups readers also look at.
AIHubMix