GPT-5.6 Luna vs Kimi K3
GPT-5.6 Luna (OpenAI) and Kimi K3 (Moonshot) compared on benchmarks, pricing, context window, and use-case rankings.
- Kimi K3 leads on overall intelligence (ECI 157 vs 156).
- GPT-5.6 Luna is cheaper on input tokens — $0.20 vs $2.80 per 1M.
- GPT-5.6 Luna has a larger context window (1.1M vs 1M tokens).
- Kimi K3 ranks higher for coding (99th vs 76th percentile).
| Attribute | GPT-5.6 Luna | Kimi K3 |
|---|---|---|
| Scores | ||
| Intelligence (ECI) | 156 | 157 |
| Coding | 76 | 99 |
| Math | 87 | 81 |
| Reasoning & Knowledge | 57 | 78 |
| Agentic & Tools | — | 84 |
| Specifications | ||
| Developer | OpenAI | Moonshot |
| Family | GPT | Kimi |
| Released | Jul 9, 2026 | Jul 16, 2026 |
| Parameters | — | 2.8T |
| Availability | API access | Open weights (non-commercial) |
| Context window | 1.1M | 1M |
| Price — $/M input | $0.20 | $2.80 |
| Price — $/M output | $1.20 | $14.00 |
| Inputs | file, image, text | text, image, video |
| Outputs | text | text |
| Benchmarks | ||
| AIME 2024/2025 | 98% | 97% |
| ARC-AGI | 88% | 95% |
| ARC-AGI-2 | 60% | 60% |
| GPQA Diamond | 92% | 93% |
| Humanity's Last Exam | 40% | 47% |
| SciCode | 53% | 59% |
| SimpleBench | 47% | — |
| SimpleQA Verified | 42% | 43% |
| WebDev Arena | 1523 | 1682 |
| WeirdML | 61% | 83% |
| APEX | — | 39% |
Use-case scores are 0–100 percentile composites across each area’s benchmarks, ranked against every model from the past year. Highlighted cells lead each row. Open a model for the full picture.
Frequently asked questions
Is GPT-5.6 Luna better than Kimi K3?
On Epoch AI's Capabilities Index, Kimi K3 scores higher (157) than GPT-5.6 Luna (156). The right pick depends on your task — compare their coding, math, and reasoning scores in the table above.
Which is cheaper, GPT-5.6 Luna or Kimi K3?
GPT-5.6 Luna is cheaper on input tokens at $0.20 per million, versus $2.80 (representative OpenRouter pricing).
Which has a larger context window, GPT-5.6 Luna or Kimi K3?
GPT-5.6 Luna supports up to 1.1M tokens, compared with 1M for the other.
Which is better for coding, GPT-5.6 Luna or Kimi K3?
Across coding benchmarks like SWE-bench Verified and Terminal-Bench, Kimi K3 ranks higher — 99th vs 76th percentile among the models tracked on Model Beat.
Want a different match-up? Open the compare tool to add or swap models.
More comparisons
Benchmarks & model data from Epoch AI (CC BY); pricing & specs from OpenRouter. ECI = Epoch Capabilities Index.