Benchmark results

The charts compare KODA-35B with the Qwen3.6 base model and other models on KernelBench and SOL-ExecBench. The horizontal axis shows total parameter count on a logarithmic scale. The red star marks KODA-35B.
Best values are bold, including ties. Compared with its Qwen3.6 base model, KODA-35B improves correctness by 35.2 percentage points on KernelBench and 31.1 points on SOL-ExecBench, based on the displayed scores. It ties for the highest KernelBench correctness and reaches the second-highest KernelBench median speedup in this comparison.
KernelBench (270 problems)
Table with columns: Model, Correctness, Fast@1, Fast@2, Median speedup| Model | Correctness | Fast@1 | Fast@2 | Median speedup |
|---|
| KODA-35B | 95.6% | 88.9% | 61.9% | 3.15× |
| Qwen3.6-35B-A3B | 60.4% | 24.1% | 10.7% | 0.94× |
| Qwen3.8-Flash-Next | 94.8% | 82.2% | 51.9% | 2.27× |
| DeepSeek-V4.1-Flash-1017 | 92.6% | 90.0% | 70.4% | 4.31× |
| DeepSeek-V4-Flash-0731 | 95.6% | 77.4% | 52.6% | 2.26× |
| GLM-5.3 | 94.1% | 84.4% | 62.6% | 2.97× |
| GLM-5.3-Flash | 95.6% | 80.0% | 48.1% | 2.02× |
| Kimi-K3 | 93.7% | 84.8% | 64.1% | 2.99× |
| MiMo-v2.6-Flash-MOPD | 89.6% | 71.1% | 39.6% | 1.72× |
| Recursive-Flint-40B | 68.9% | 18.1% | 9.3% | 0.48× |
| MusaCoder-27B | 73.0% | 26.7% | 14.1% | 0.95× |
| Kevin-32B | 5.2% | 1.1% | 0.7% | 0.64× |
| IndustrialCoder-32B-Thinking | 23.7% | 3.0% | 1.1% | 0.27× |
SOL-ExecBench (202 problems with fixed workload sampling)
Table with columns: Model, Correctness, Fast@1, Fast@2, Median speedup| Model | Correctness | Fast@1 | Fast@2 | Median speedup |
|---|
| KODA-35B | 95.5% | 78.7% | 48.0% | 2.10× |
| Qwen3.6-35B-A3B | 64.4% | 17.8% | 8.9% | 0.73× |
| Qwen3.8-Flash-Next | 95.5% | 74.8% | 44.1% | 1.70× |
| DeepSeek-V4.1-Flash-1017 | 91.6% | 81.7% | 54.5% | 3.15× |
| DeepSeek-V4-Flash-0731 | 98.0% | 67.3% | 41.1% | 1.37× |
| GLM-5.3 | 93.6% | 78.7% | 50.0% | 2.28× |
| GLM-5.3-Flash | 96.0% | 72.8% | 43.6% | 1.57× |
| Kimi-K3 | 94.6% | 78.2% | 44.6% | 1.81× |
| MiMo-v2.6-Flash-MOPD | 82.2% | 57.4% | 32.7% | 1.43× |
| Recursive-Flint-40B | 65.8% | 13.4% | 8.4% | 0.30× |
| MusaCoder-27B | 71.3% | 26.7% | 14.9% | 0.87× |
| Kevin-32B | 2.5% | 0.0% | 0.0% | 0.74× |
| IndustrialCoder-32B-Thinking | 25.7% | 3.5% | 2.0% | 0.59× |
Usage
Serve KODA with a compatible inference engine such as vLLM or SGLang, and connect its model API to the KODA agent. The agent compiles and evaluates candidates in managed sandboxes and feeds the results back to the model.
Model serving and kernel evaluation use separate GPU resources. Refer to the agent repository for configuration and evaluation instructions.
Model provenance
This repository provides the KODA model weights, tokenizer, and configuration in Hugging Face format. KODA is fine-tuned from Qwen3.6-35B-A3B.
License
The model is distributed under Apache License 2.0. See LICENSE for the license text and upstream notices. We acknowledge the Qwen team for the base model. The agent code and benchmark datasets have their own licenses.