token.app › Compare › Kimi K3 vs GPT-5.6 Luna
Kimi K3 vs GPT-5.6 Luna
GPT-5.6 Luna leads on 3 of 5 shared benchmarks. GPT-5.6 Luna is 27× cheaper per token on a blended 3:1 basis. GPT-5.6 Luna accepts the larger context window (1.1M).
Head to head
Blended price uses a 3:1 input:output mix
| Kimi K3 | GPT-5.6 Luna | |
|---|---|---|
| Provider | Moonshot AI | OpenAI |
| Input $/1M | $3.00 | $0.100 |
| Output $/1M | $15.00 | $0.600 |
| Blended $/1M | $6.00 | $0.225 |
| Context window | 1.0M | 1.1M |
| Max output | — | 128K |
| Released | Jul 2026 | Jul 2026 |
| Licence | Open weights | Proprietary |
| GPQA Diamond | 93.1% | 91.6% |
| FrontierMath T1–3 | 72.2% | 82.1% |
| FrontierMath T4 | 39.0% | 61.0% |
| SimpleQA Verified | 42.7% | 41.7% |
| OTIS Mock AIME | 97.2% | 98.3% |
Benchmark scores independently run and published by Epoch AI, 'AI Benchmarking Hub', used under CC BY 4.0. Only benchmarks with a score for BOTH models are compared.
Choose Kimi K3 if…
- Higher GPQA Diamond (93.1% vs 91.6%)
- Higher SimpleQA Verified (42.7% vs 41.7%)
- Open weights — self-hostable
- Newer model (released Jul 2026)
Choose GPT-5.6 Luna if…
- Costs less per token — $0.225 vs $6.00 blended
- Higher FrontierMath T1–3 (82.1% vs 72.2%)
- Higher FrontierMath T4 (61.0% vs 39.0%)
- Higher OTIS Mock AIME (98.3% vs 97.2%)
- Larger context window (1.1M)
FAQ
Which is better, Kimi K3 or GPT-5.6 Luna?
Across the 5 benchmarks both models have independently-run scores for, Kimi K3 leads on 2 and GPT-5.6 Luna leads on 3. Benchmarks are one input — price and context window are on this page too.
Is Kimi K3 cheaper than GPT-5.6 Luna?
GPT-5.6 Luna is cheaper: $0.225 versus $6.00 per million tokens blended at 3:1 input:output. Input alone: $3.00 vs $0.100. Output alone: $15.00 vs $0.600.
What context windows do Kimi K3 and GPT-5.6 Luna have?
Kimi K3: 1,048,576 tokens. GPT-5.6 Luna: 1,050,000 tokens.
Who makes Kimi K3 and GPT-5.6 Luna?
Kimi K3 is made by Moonshot AI. GPT-5.6 Luna is made by OpenAI.