Kotlin-bench: leaderboard
Kotlin-specific code generation benchmark from Firebender evaluating LLM proficiency in Kotlin programming tasks.
Metric: Pass Rate (%). Source: firebender.com. Saturation forecast: Around January 2027. 17 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.3 Codex | 49 | #69 |
| 2 | Claude Opus 4.5 | 46 | #83 |
| 3 | GPT-5.2 | 43 | #110 |
| 4 | Gemini 3 Pro | 42 | #80 |
| 5 | Gemini 3 Flash | 41 | #99 |
| 6 | GPT-5.1 Codex | 41 | #159 |
| 7 | Claude Sonnet 4.5 | 37 | #142 |
| 8 | GPT-5.2 Codex | 37 | #95 |
| 9 | Gemini 3.1 Pro (Preview) | 36 | #58 |
| 10 | Kimi K2.5 | 31 | #144 |
| 11 | GLM-5 | 30 | #141 |
| 12 | Claude Sonnet 4 | 29 | #199 |
| 13 | GLM-4.7 | 22 | #192 |
| 14 | GLM-4.7 FP8 | 15 | #209 |
| 15 | GPT-4.1 | 12 | #248 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=kotlin-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.