SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.17568 · 2023

CodeGeeX: A Pre-Trained Model for Code Generation with Multilingual Benchmarking on HumanEval-X

arXiv · PDF · Open in the Atlas

Code that ran

6 of 8 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 4 have no contract verdict; 2 failures of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2303.17568.json · badge: /api/ran/2303.17568.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
THUDM/CodeGeeX canonical (classifier label)
classifier basis not recorded on the link
2 ran · 0 no recorded run
thudm/codegeex2 community (classifier label)
classifier basis not recorded on the link
4 ran · 2 no recorded run
FunctionStatusWhere it lives
add_code_generation_args Ran
ran on a synthesized input
thudm/codegeex2/demo/fastapicpu.py
code served (permissive licence) · get_code("42a53b2c340cdf87")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ArgumentParser. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import argparse

ARGS = {
    'parser': argparse.ArgumentParser(description='CodeGeeX2 Demo')
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
add_code_generation_args Ran
ran on a synthesized input
thudm/codegeex2/demo/run_demo.py
code served (permissive licence) · get_code("c25730f1da3ae813")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ArgumentParser. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import argparse

ARGS = {
    'parser': argparse.ArgumentParser(description='CodeGeeX2 Demo')
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
auto_configure_device_map Ran
ran on a synthesized input
thudm/codegeex2/demo/gpus.py
code served (permissive licence) · get_code("cd91456d18781855")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'num_gpus': 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
fast_gelu Ran
ran; our contract draft was wrong, not the code
THUDM/CodeGeeX/codegeex/torch/codegeex_model.py
code served (permissive licence) · get_code("2a6537281349742d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function takes a single parameter 'x' which is used in mathematical operations
# It should be a tensor that supports element-wise operations like abs, exp, division, etc.
x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
postprocess_generation Ran
ran on a synthesized input
thudm/codegeex2/evaluation/evaluation.py
code served (permissive licence) · get_code("637b4d55eb53b273")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

# Test case 1: completion mode with markdown code blocks
sample_1 = {
    'generation': 'Here is some code:\n```python\nprint("hello")\n```\nEnd of code.'
}

# Test case 2: instruction mode with markdown code blocks
sample_2 = {
    'generation': 'Here is the solution:\n```python\nprint("world")\n```\nDone.'
}

# Test case 3: instruction mode without markdown code blocks
sample_3 = {
    'generation': 'This is plain text without code blocks.'
}

# Test case 4: completion mode (default)
sample_4 = {
    'generation': 'def foo():\n    return 42'
}

ARGS = {
    'sample': sample_1,
    'generation_mode': 'completion'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
set_weight_decay Ran
ran; our contract draft was wrong, not the code
THUDM/CodeGeeX/codegeex/mindspore/finetune.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from yingweima2022/codellm, is pointer only (licence: NONE) · get_code("f12f9486ba2960e6")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from yingweima2022/codellm, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_model Not yet run
unverified (harvested, no recorded run)
thudm/codegeex2/demo/run_demo.py
code served (permissive licence) · get_code("b5f56a90751fcf4d")
How it ran

no invocation recorded.

load_model_on_gpus Not yet run
unverified (harvested, no recorded run)
thudm/codegeex2/demo/gpus.py
code served (permissive licence) · get_code("eb072eed78a57ca7")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 5 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
set_weight_decay 3 other papers
3 later
2309.02772 (2023-09)
Hot or Cold? Adaptive Temperature Sampling for Code Generation with Large Language Models
2309.16298 (2023-09)
At Which Training Stage Does Code Data Help LLMs Reasoning?
aaai_27798 (2024 (year only))
Hot or Cold? Adaptive Temperature Sampling for Code Generation with Large Language Models
fast_gelu 2 other papers
2 later
2309.02772 (2023-09)
Hot or Cold? Adaptive Temperature Sampling for Code Generation with Large Language Models
aaai_27798 (2024 (year only))
Hot or Cold? Adaptive Temperature Sampling for Code Generation with Large Language Models
load_model_on_gpus 2 other papers
2 later
2306.04188 (2023-06)
A New Dataset and Empirical Study for Sentence Simplification in Chinese
2406.12793 (2024-06)
ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Large pre-trained code generation models, such as OpenAI Codex, can generate syntax- and function-correct code, making the coding of programmers more productive and our pursuit of artificial general intelligence closer. In this paper, we introduce CodeGeeX, a multilingual model with 13 billion parameters for code generation. CodeGeeX is pre-trained on 850 billion tokens of 23 programming languages as of June 2022. Our extensive experiments suggest that CodeGeeX outperforms multilingual code models of similar scale for both the tasks of code generation and translation on HumanEval-X. Building upon HumanEval (Python only), we develop the HumanEval-X benchmark for evaluating multilingual models by hand-writing the solutions in C++, Java, JavaScript, and Go. In addition, we build CodeGeeX-based extensions on Visual Studio Code, JetBrains, and Cloud Studio, generating 4.7 billion tokens for tens of thousands of active users per week. Our user study demonstrates that CodeGeeX can help to increase coding efficiency for 83.4% of its users. Finally, CodeGeeX is publicly accessible and in Sep. 2022, we open-sourced its code, model weights (the version of 850B tokens), API, extensions, and HumanEval-X at https://github.com/THUDM/CodeGeeX.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2303.17568")
get_code_for_paper("2303.17568")
have("2303.17568")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2303.17568.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2303.17568.svg)](https://syntology.ai/paper/2303.17568)

Connect an agent — have() is free.