Paper 2603.16790
InCoder-32B: Code Foundation Model for Industrial Scenarios
- Published
- Mar 2026
- Research lab
- Independent
- Citations
- 6
- GitHub
- Not linked
01 In brief
Summary
InCoder-32B is a 32B-parameter code foundation model designed to unify code intelligence across general programming and industrial domains such as chip design, GPU kernel optimization, embedded systems, compiler optimization, and 3D modeling.
It is trained from scratch using a three-stage Code-Flow pipeline: pre-training with curated industrial data, mid-training that extends context from 8K to 128K tokens with synthetic reasoning data, and post-training with execution-grounded verification.
The model is evaluated on 14 general and 9 industrial benchmarks, achieving competitive results on general tasks (e.g., 74.8% on SWE-bench Verified, 49.14% on LiveCodeBench) and establishing strong open-source baselines on industrial benchmarks, including leading performance on RealBench, CAD-Coder, and KernelBench.
Error analysis reveals common failures in syntax, API usage, and functional correctness, while scaling SFT data improves performance across most benchmarks.
The work highlights the need for domain-specific training to bridge the gap between general code LLMs and industrial requirements.
02 From the paper
Abstract
Recent code large language models have achieved remarkable progress on general programming tasks. Nevertheless, their performance degrades significantly in industrial scenarios that require reasoning about hardware semantics, specialized language constructs, and strict resource constraints. To address these challenges, we introduce InCoder-32B (Industrial-Coder-32B), the first 32B-parameter code foundation model unifying code intelligence across chip design, GPU kernel optimization, embedded systems, compiler optimization, and 3D modeling. By adopting an efficient architecture, we train InCoder-32B from scratch with general code pre-training, curated industrial code annealing, mid-training that progressively extends context from 8K to 128K tokens with synthetic industrial reasoning data, and post-training with execution-grounded verification. We conduct extensive evaluation on 14 mainstream general code benchmarks and 9 industrial benchmarks spanning 4 specialized domains. Results show InCoder-32B achieves highly competitive performance on general tasks while establishing strong open-source baselines across industrial domains.