SWE-Lancer (IC-Diamond subset)

Paper

Progress Over Time

Interactive timeline showing model performance evolution on SWE-Lancer (IC-Diamond subset)

State-of-the-art frontier
Open
Proprietary

SWE-Lancer (IC-Diamond subset) Leaderboard

6 models
ContextCostLicense
1
OpenAI
OpenAI
2400K$1.75 / $14.00
3
OpenAI
OpenAI
400K$1.75 / $14.00
4
OpenAI
OpenAI
5
OpenAI
OpenAI
128K$2.50 / $10.00
6
OpenAI
OpenAI
Notice missing or incorrect data?
About this benchmark

What is SWE-Lancer (IC-Diamond subset)?

SWE-Lancer (IC-Diamond subset) is a benchmark of real-world freelance software engineering tasks from Upwork, ranging from $50 bug fixes to $32,000 feature implementations. It evaluates AI models on independent engineering tasks using end-to-end tests triple-verified by experienced software engineers, and includes managerial tasks where models choose between technical implementation proposals.

SWE-Lancer (IC-Diamond subset) is a text benchmark evaluating models on reasoning and code tasks. LLM Stats tracks 6 models on this benchmark, scored on a 0–1 scale. The current average is 0.5, with the leader at 1.0.

Compare leaders on the best AI for reasoning and best AI for code leaderboards.

Current leaders

GPT-5 from OpenAI currently leads the SWE-Lancer (IC-Diamond subset) leaderboard with a score of 1.000 across 6 evaluated AI models.

1GPT-5OpenAI100.0%
2GPT-5.3 CodexOpenAI81.4%
3GPT-5.2OpenAI74.6%

Source paper

Title
SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering?
Authors
Samuel Miserendino, Michele Wang, Tejal Patwardhan, Johannes Heidecke
Published
Abstract

We introduce SWE-Lancer, a benchmark of over 1,400 freelance software engineering tasks from Upwork, valued at \$1 million USD total in real-world payouts. SWE-Lancer encompasses both independent engineering tasks--ranging from \$50 bug fixes to \$32,000 feature implementations--and managerial tasks, where models choose between technical implementation proposals. Independent tasks are graded with end-to-end tests triple-verified by experienced software engineers, while managerial decisions are assessed against the choices of the original hired engineering managers. We evaluate model performance and find that frontier models are still unable to solve the majority of tasks. To facilitate future research, we open-source a unified Docker image and a public evaluation split, SWE-Lancer Diamond (https://github.com/openai/SWELancer-Benchmark). By mapping model performance to monetary value, we hope SWE-Lancer enables greater research into the economic impact of AI model development.

FAQ

Common questions about the SWE-Lancer (IC-Diamond subset) benchmark and leaderboard.

What is the SWE-Lancer (IC-Diamond subset) benchmark?

SWE-Lancer (IC-Diamond subset) is a benchmark of real-world freelance software engineering tasks from Upwork, ranging from $50 bug fixes to $32,000 feature implementations. It evaluates AI models on independent engineering tasks using end-to-end tests triple-verified by experienced software engineers, and includes managerial tasks where models choose between technical implementation proposals.

What is the SWE-Lancer (IC-Diamond subset) leaderboard?

The SWE-Lancer (IC-Diamond subset) leaderboard ranks 6 AI models based on their performance on this benchmark. Currently, GPT-5 by OpenAI leads with a score of 1.000. The average score across all models is 0.489.

What is the highest SWE-Lancer (IC-Diamond subset) score?

The highest SWE-Lancer (IC-Diamond subset) score is 1.000, achieved by GPT-5 from OpenAI.

How many models are evaluated on SWE-Lancer (IC-Diamond subset)?

6 models have been evaluated on the SWE-Lancer (IC-Diamond subset) benchmark, with 0 verified results and 6 self-reported results.

Where can I find the SWE-Lancer (IC-Diamond subset) paper?

The SWE-Lancer (IC-Diamond subset) paper is available at https://arxiv.org/abs/2502.12115. The paper details the methodology, dataset construction, and evaluation criteria.

What categories does SWE-Lancer (IC-Diamond subset) cover?

SWE-Lancer (IC-Diamond subset) is categorized under reasoning and code. The benchmark evaluates text models.

How recent are the SWE-Lancer (IC-Diamond subset) leaderboard results?

The SWE-Lancer (IC-Diamond subset) leaderboard was last updated in July 2026 and currently includes 6 evaluated models.