Kimi K3 (low)
AI model by Kimi. Real-time pricing and benchmark data.
Benchmarks
Editorial Analysis
Profile. Kimi K3 (low) is a premium-tier model from Kimi at $15.00/M output tokens. Benchmark scores (e.g. GPQA Diamond 84.2%) suggest it targets workloads where correctness matters more than cost — long-running agents, code generation at scale, and complex multi-step reasoning.
Latency. On latency, Kimi K3 (low) reports 32 tok/s output speed and 3352ms time-to-first-token. Output speed is on the slower end; for real-time chat consider pairing with a streaming front-end or routing latency-sensitive requests to a faster model.
Cost model. For a workload of 1M input tokens + 100K output tokens, Kimi K3 (low) costs $4.50 per million input-equivalent requests. Multiply by your monthly request volume to project the bill, and compare against the speed/correctness profile above to decide if the trade-off is worth it for your workload.
Editorial summary generated from public pricing and benchmark data (Artificial Analysis API). All numbers are sourced — see the Artificial Analysis methodology for benchmark definitions. Last refreshed 2026-07-31.
Compare with similar models
| Model | Input | Output | Speed |
|---|---|---|---|
Kimi K3 (low)Current | $3.00 | $15.00 | 32 tok/s |
| Claude Sonnet 4.6 (Non-reasoning, Low Effort) | $3.00 | $15.00 | 47 tok/s |
| GPT-4o (May '24) | $5.00 | $15.00 | 0 tok/s |
| GPT-5.4 (xhigh) | $2.50 | $15.00 | 0 tok/s |
| GPT-5.4 (low) | $2.50 | $15.00 | 0 tok/s |
| GPT-5.4 (Non-reasoning) | $2.50 | $15.00 | 0 tok/s |