GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
InternScience
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
InternScience
Model Tree
Input Modalities
Output Modalities
Supported Functionality
We primarily evaluate our trained models on MLE-Bench Lite, a 22-task subset of MLE-Bench consisting of historical Kaggle competitions for end-to-end machine learning engineering. We report the percentage of tasks awarded each medal tier (Gold, Silver, and Bronze), the percentage receiving any medal (Any Medal), and the percentage exceeding the median human score (Med+), with Any Medal as the primary metric. All five metrics are computed over the 22 tasks and reported as percentages.
The first five metric columns use the MLExpert Harness; the last two columns report the general-purpose harness and its medal rate. All values are percentages.
MLExpert achieves the highest medal rate among same-scale models and reaches the performance level of larger frontier models. Under the unified MLExpert Harness, MLExpert raises the medal rate of its Qwen3.6-35B-A3B base model from 37.9% to 72.7%, achieving the highest rate among the same-scale models evaluated and performance comparable to GLM-5.3 and Kimi-K3. These results show that long-horizon agentic post-training substantially improves the base model's MLE capability, bringing a 35B model to a performance level comparable to larger frontier models.
MLExpert retains its advantage over the base model under a general-purpose coding harness. With Claude Code, MLExpert achieves a 69.7% medal rate, compared with 42.4% for its base model, and exceeds the strongest evaluated same-scale baseline by 22.7 percentage points. The consistent advantage over the base model across both harnesses suggests that the improvement from post-training is not specific to the MLExpert-Harness and remains effective in a general-purpose agentic coding environment.
We also evaluate on NatureBench, following the ten-task Lite subset used in Frontis-MA1. M denotes Match-SOTA (g ≥ 0), while S denotes Surpass-SOTA (g > 0.1), which serves as a key metric for directly assessing agents' ability to surpass published scientific results.
On NatureBench Lite, MLExpert improves scientific task performance over the base model. With Claude Code, MLExpert raises the Surpass-SOTA rate from 0.0% to 50.0%, exceeding Nex-N2.5-mini at 20.0% and matching GLM-5.3 at 50.0%. Its Match-SOTA rate also improves from 40.0% to 70.0%. These results show that post-training improves scientific task optimization, enabling MLExpert to match or exceed published reference results on more tasks.
The commands below load InternScience/MLExpert and serve an API endpoint at http://localhost:8000/v1.
Install SGLang with uv:
shell
uv venv --python 3.12 --seed --managed-pythonsource .venv/bin/activateuv pip install sglangText Generation
shell
python -m sglang.launch_server \ --model-path InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tp-size 2 \ --context-length 262144 \ --reasoning-parser qwen3 \ --language-model-onlyTool Use
shell
python -m sglang.launch_server \ --model-path InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tp-size 2 \ --context-length 262144 \ --reasoning-parser qwen3 \ --tool-call-parser qwen3_coder \ --language-model-onlyInstall vLLM with uv:
shell
uv venv --python 3.12 --seed --managed-pythonsource .venv/bin/activateuv pip install vllm --torch-backend=autoText Generation
shell
vllm serve InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tensor-parallel-size 2 \ --max-model-len 262144 \ --reasoning-parser qwen3 \ --language-model-onlyTool Calling
shell
vllm serve InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tensor-parallel-size 2 \ --max-model-len 262144 \ --reasoning-parser qwen3 \ --enable-auto-tool-choice \ --tool-call-parser qwen3_coder \ --language-model-only| Parameter | Value |
|---|---|
temperature | 1.0 |
top_p | 0.95 |
top_k | 20 |
min_p | 0.0 |
presence_penalty | 1.5 |
repetition_penalty | 1.0 |
These parameters apply to both thinking and non-thinking modes. Set chat_template_kwargs.enable_thinking to true to enable thinking or false to disable it. Machine-readable profiles are provided in inference_profiles.json.
uv venv --python 3.12 --seed --managed-pythonsource .venv/bin/activateuv pip install sglangpython -m sglang.launch_server \ --model-path InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tp-size 2 \ --context-length 262144 \ --reasoning-parser qwen3 \ --language-model-onlypython -m sglang.launch_server \ --model-path InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tp-size 2 \ --context-length 262144 \ --reasoning-parser qwen3 \ --tool-call-parser qwen3_coder \ --language-model-onlyuv venv --python 3.12 --seed --managed-pythonsource .venv/bin/activateuv pip install vllm --torch-backend=autovllm serve InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tensor-parallel-size 2 \ --max-model-len 262144 \ --reasoning-parser qwen3 \ --language-model-onlyvllm serve InternScience/MLExpert \ --host 127.0.0.1 \ --port 8000 \ --tensor-parallel-size 2 \ --max-model-len 262144 \ --reasoning-parser qwen3 \ --enable-auto-tool-choice \ --tool-call-parser qwen3_coder \ --language-model-only