Model comparison
GPT-5.6 Luna vs Kimi K2
Head-to-head evidence from 11 shared benchmark results across 4 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: GPT-5.6 Luna #22 (Estimated); Kimi K2 #189 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.6 Luna and Kimi K2 share 11 comparable benchmark results. 1 of 8 categories are comparable. 30 results are unique to GPT-5.6 Luna; 3 to Kimi K2.
Updated July 23, 2026- Shared results
- 11
- GPT-5.6 Luna only
- 30
- Kimi K2 only
- 3
- Comparable categories
- 1 / 8
Pick GPT-5.6 Luna if you want the stronger benchmark profile. Kimi K2 only becomes the better choice if you want the cheaper token bill or you would rather avoid the extra latency and token burn of a reasoning model.
Confidence note. This is a partial-evidence comparison with 11 shared benchmark results across 4 evidence categories; 1 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
GPT-5.6 Luna is clearly ahead on the BenchAlign aggregate, 67.17 to 27.19. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.6 Luna's sharpest advantage is in mathematics, where it averages 73.6 against 16.1. The single biggest benchmark swing on the page is FrontierMath v2 (Tier 4), 58.500% to 0.000%.
GPT-5.6 Luna is also the more expensive model on tokens at $1.00 input / $6.00 output per 1M tokens, versus $0.60 input / $2.50 output per 1M tokens for Kimi K2. That is roughly 2.4x on output cost alone. GPT-5.6 Luna is the reasoning model in the pair, while Kimi K2 is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. GPT-5.6 Luna gives you the larger context window at 1M, compared with 128K for Kimi K2.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | GPT-5.6 Luna | Δ | Kimi K2 |
|---|---|---|---|
| Math | GPT-5.6 Luna73.6 | Margin← 57.5 | Kimi K216.1 |
| Agentic | GPT-5.6 Luna84.1 | MarginNo overlap | Kimi K2Not measured |
| Coding | GPT-5.6 Luna62.7 | MarginNo overlap | Kimi K2Not measured |
| Knowledge | GPT-5.6 Luna92.3 | MarginNo overlap | Kimi K2Not measured |
| Multimodal | GPT-5.6 Luna78.4 | MarginNo overlap | Kimi K2Not measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
FrontierMath v2 (Tier 4)
MathA 58.500%B 0.000%Winner: GPT-5.6 LunaΔ 58.5FrontierMath v2 (Tier 4): GPT-5.6 Luna scored 58.500%; Kimi K2 scored 0.000%. GPT-5.6 Luna wins this benchmark. - Source ↗
FrontierMath v2 (Tiers 1-3)
MathA 78.600%B 21.404%Winner: GPT-5.6 LunaΔ 57.2FrontierMath v2 (Tiers 1-3): GPT-5.6 Luna scored 78.600%; Kimi K2 scored 21.404%. GPT-5.6 Luna wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.6 Luna | Kimi K2 | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.6 Luna$1 input / $6 output | Kimi K2$0.6 input / $2.5 output | Kimi K2 has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.6 LunaNot available | Kimi K243 tok/s | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | GPT-5.6 LunaNot available | Kimi K21.51 s | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | GPT-5.6 Luna1M | Kimi K2128K | GPT-5.6 Luna lists the larger context window. |
Benchmark Deep Dive
Agentic16 benchmarks
| Benchmark | GPT-5.6 Luna | Kimi K2 | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 84.7% | — | Not comparable |
| BrowseCompSource | 83.3% | — | Not comparable |
| OSWorld 2.0Source | 45.6% | — | Not comparable |
| CyberGymSource | 77.9% | — | Not comparable |
| ExploitGymSource | 12.4% | — | Not comparable |
| ToolathlonSource | 53.4% | — | Not comparable |
| AA Agentic IndexSource | 45.6% | — | Not comparable |
| GDPval-AASource | 54.2% | — | Not comparable |
| GDPval-AASource | 1584 | — | Not comparable |
| AA Harvey LABSource | 87.9% | — | Not comparable |
| AA ITBenchSource | 40.3% | — | Not comparable |
| AA Tau3 BankingSource | 27.2% | — | Not comparable |
| AA AutomationBenchSource | 42.2% | — | Not comparable |
| aaTerminalBench21Source | 80.9% | — | Not comparable |
| APEX-Agents-AASource | 35.8% | — | Not comparable |
| τ²-bench resultsSource | — | 61.1% | Not comparable |
Coding7 benchmarks
| Benchmark | GPT-5.6 Luna | Kimi K2 | Result |
|---|---|---|---|
| SWE-bench ProSource | 62.7% | — | Not comparable |
| Terminal-Bench 2.0Source | 84.7% | — | Not comparable |
| deepSweSource | 67.2% | — | Not comparable |
| FrontierCode 1.1 ExtendedSource | 55.1% | — | Not comparable |
| cursorBench32Source | 61.1% | — | Not comparable |
| AA Coding IndexSource | 71.5% | — | Not comparable |
| AA-SciCodeSource | 52.5% | 34.5% | GPT-5.6 Luna leads |
Reasoning3 benchmarks
Knowledge10 benchmarks
| Benchmark | GPT-5.6 Luna | Kimi K2 | Result |
|---|---|---|---|
| GPQASource | 92.3% | — | Not comparable |
| GPQA-DSource | 92.3% | — | Not comparable |
| HealthBench ProfessionalSource | 55.7% | — | Not comparable |
| HealthBench HardSource | 32.0% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 51.2% | 19.4% | GPT-5.6 Luna leads |
| AA-GPQA DiamondSource | 91.1% | 76.6% | GPT-5.6 Luna leads |
| AA-HLESource | 37.2% | 7.0% | GPT-5.6 Luna leads |
| AA-Omniscience IndexSource | -11.2% | -27.5% | GPT-5.6 Luna leads |
| AA-Omniscience AccuracySource | 41.5% | 26.8% | GPT-5.6 Luna leads |
| AA-Omniscience Hallucination RateSource | 90.1% | 74.2% | Kimi K2 leads |
MathGPT-5.6 Luna wins3 benchmarks
Multimodal4 benchmarks
Inst. Following1 benchmarks
| Benchmark | GPT-5.6 Luna | Kimi K2 | Result |
|---|---|---|---|
| AA-IFBenchSource | — | 41.5% | Not comparable |
Frequently Asked Questions (2)
Which is better, GPT-5.6 Luna or Kimi K2?
GPT-5.6 Luna is ahead on BenchLM's BenchAlign leaderboard, 67.17 to 27.19. The biggest single separator in this matchup is FrontierMath v2 (Tier 4), where the scores are 58.500% and 0.000%.
Which is better for math, GPT-5.6 Luna or Kimi K2?
GPT-5.6 Luna has the edge for math in this comparison, averaging 73.6 versus 16.1. Inside this category, FrontierMath v2 (Tier 4) is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.