# MLExpert

> MLExpert is a 35B Mixture-of-Experts agentic model for autonomous machine learning engineering. It organizes experiments, writes and revises code, and improves solutions through sustained interaction with execution environments.

- **Organisation:** InternScience
- **Category:** Chat & Text Generation, Reasoning Models
- **Parameters:** 34.7B total, 3.5B active (MoE)
- **Context length:** 262,144 tokens
- **Licence:** Apache 2.0 (Commercial use OK)
- **Released:** 2026-10-01
- **Hugging Face:** 3.4K downloads (30 days), 17 likes

## Estimated memory

| Quant | ≈ GB |
|---|---|
| Q2_K | 12.27 |
| Q3_K_M | 17.69 |
| Q4_K_M | 21.67 |
| Q5_K_M | 25.23 |
| Q6_K | 29.05 |
| Q8_0 | 37.14 |
| F16 | 68.91 |

## Runs with

- [llama.cpp](https://selfhost.directory/project/llamacpp)
- [SGLang](https://selfhost.directory/project/sglang)
- [Transformers](https://github.com/huggingface/transformers)
- [vLLM](https://selfhost.directory/project/vllm)
- [Jan](https://jan.ai) (community)
- [KoboldCpp](https://selfhost.directory/project/koboldcpp) (community)
- [LM Studio](https://lmstudio.ai) (community)
- [LocalAI](https://selfhost.directory/project/localai) (community)
- [Ollama](https://selfhost.directory/project/ollama) (community)

## About

MLExpert is a 35B Mixture-of-Experts agentic model for autonomous machine learning engineering. It organizes experiments, writes and revises code, and improves solutions through sustained interaction with execution environments. Hierarchical agentic SFT develops general tool-use and domain-specific optimization skills, while Windowed Interaction Reinforcement Learning (WIRL) trains multi-turn decisions using feedback from completed experiments.

• Long-horizon optimization: Plans experiments, interprets execution feedback, revisits previous solutions, and selects submission candidates within a persistent workspace.
• Strong MLE performance: Achieves a 72.7% medal rate on MLE-Bench Lite, matching GLM-5.3 and Kimi-K3 under the MLExpert Harness.
• General-purpose harness compatibility: Achieves a 69.7% medal rate with Claude Code, compared with 42.4% for its base model.
• Scientific task optimization: Matches published scientific results on seven of ten NatureBench Lite tasks and surpasses them on five.

## Alternatives

- [ornith-ai/Ornith-1.0-35B](https://selfhost.directory/models/ornith-ai/ornith-1.0-35b) (35B (3.8B active), MIT, ≈21.87 GB at Q4): Aloha! 🌺 Today, we are releasing Ornith-1.0, a self-improving family of open-source models for agentic coding.
- [nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16](https://selfhost.directory/models/nvidia/nvidia-nemotron-3.5-lightning-30b-a3b-bf16) (31.6B (3B active), openmdw-1.1, ≈19.63 GB at Q4): NVIDIA Nemotron™ is a family of open models with open weights, training data, and recipes, delivering leading efficiency and accuracy for building specialized AI agents.
- [farbodtavakkoli/OTel-2.0-LLM-31B-IT](https://selfhost.directory/models/farbodtavakkoli/otel-2.0-llm-31b-it) (31.3B, Apache 2.0, ≈26.54 GB at Q4): Versioning notice: This card describes the current OTel 2.0 OSFT checkpoint. The model weights may be updated over time. For reproducible evaluation or production deployment, pin a specific model revision, checkpoint hash, or release tag rather than relying on the floating latest version.
- [poolside/Laguna-XS-2.1](https://selfhost.directory/models/poolside/laguna-xs-2.1) (33.4B (2.2B active), Custom licence, ≈21.58 GB at Q4): Laguna XS 2.1 is a 33B total parameter Mixture-of-Experts model with 3B activated parameters per token designed for agentic coding and long-horizon work on a local machine.
- [ornith-ai/Ornith-1.5-35B-A3B](https://selfhost.directory/models/ornith-ai/ornith-1.5-35b-a3b) (36B (3B active), MIT, ≈22.44 GB at Q4): Chirp Chirp! 🐦 We are introducing Ornith-1.5, a major step toward building foundation models through end-to-end self-improvement.
- [poolside/Laguna-XS.2](https://selfhost.directory/models/poolside/laguna-xs.2) (33.4B (2.2B active), Apache 2.0, ≈21.58 GB at Q4): Laguna XS.2 is a 33B total parameter Mixture-of-Experts model with 3B activated parameters per token designed for agentic coding and long-horizon work on a local machine. It uses Sliding Window Attention with per-head gating in 30 out of 40 layers for fast inference and low KV cache requirements.

More: https://selfhost.directory/models/internscience/mlexpert/alternatives

## Runs well on

- [Instinct MI50 32GB](https://selfhost.directory/models/internscience/mlexpert/on/instinct-mi50): Q5_K_M, est. 229 tok/s
- [GeForce RTX 5090](https://selfhost.directory/models/internscience/mlexpert/on/rtx-5090): Q5_K_M, est. 401 tok/s
- [Apple M4 Pro (48 GB)](https://selfhost.directory/models/internscience/mlexpert/on/apple-m4-pro-48gb): Q6_K, est. 53 tok/s
- [RTX A6000](https://selfhost.directory/models/internscience/mlexpert/on/rtx-a6000): Q8_0, est. 115 tok/s
- [Apple M1 Max (64 GB)](https://selfhost.directory/models/internscience/mlexpert/on/apple-m1-max-64gb): Q8_0, est. 60 tok/s
- [NVIDIA A100 80GB](https://selfhost.directory/models/internscience/mlexpert/on/a100-80gb): F16, est. 162 tok/s

## Similar size

- [zai-org/GLM-4.7-Flash](https://selfhost.directory/models/zai-org/glm-4.7-flash) (31.2B)
- [ornith-ai/Ornith-1.5-35B-A3B](https://selfhost.directory/models/ornith-ai/ornith-1.5-35b-a3b) (36B)
- [RecursiveSI/flint](https://selfhost.directory/models/recursivesi/flint) (39.8B)
- [Qwen/Qwen3-32B](https://selfhost.directory/models/qwen/qwen3-32b) (32.8B)
- [Qwen/Qwen3-30B-A3B](https://selfhost.directory/models/qwen/qwen3-30b-a3b) (30.5B)
- [Qwen/Qwen2.5-32B-Instruct](https://selfhost.directory/models/qwen/qwen2.5-32b-instruct) (32.8B)

## FAQ

**How much VRAM does MLExpert need?**
About 22 GB at Q4_K_M with an 8K context. Q8_0 needs about 37 GB and full precision about 69 GB. Longer contexts need more memory for the KV cache.

**Can MLExpert run on the RTX 5090?**
Yes. It fits at Q5_K_M (about 28 GB), at an estimated 401 tokens/s. It fits comfortably on the Instinct MI50 32GB, Apple M4 Pro (48 GB), RTX A6000.

**Can I use MLExpert commercially?**
Yes. It is released under Apache 2.0. Permissive: use, modify and sell freely; keep the notice and licence text.

**What is the context length of MLExpert?**
262,144 tokens (256K). The full window needs far more memory than an 8K context.

**How do I run MLExpert with Ollama or llama.cpp?**
With Ollama: ollama run hf.co/bartowski/InternScience_MLExpert-GGUF:Q4_K_M. With llama.cpp: llama-server -hf bartowski/InternScience_MLExpert-GGUF:Q4_K_M, which starts an OpenAI-compatible server.

## Links

- huggingface: https://huggingface.co/InternScience/MLExpert
- github: https://github.com/InternScience/MLExpert
- paper: https://arxiv.org/abs/2610.22344

---
Source: https://selfhost.directory/models/internscience/mlexpert