SWE-Lancer
Models complete real Upwork software tasks, from bug fixes to features, and pick between engineering proposals. Scored by end-to-end tests and payout earned.
Only the Diamond split is public. Some cards report dollars earned rather than pass rate. Since July 17th, 2025 the public repo holds 198 of the 237 IC SWE Diamond tasks. OpenAI dropped 39 tasks that failed offline, so later scores use a smaller set.
- Made by
- OpenAI
- Released
- February 2025
- Kind
- Fixing real repositories
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Code
- Tasks
- Public, plus a private test set
- Strongest model tested
- GPT-5 #70 on the capability index
- Last in a launch post
- February 2026, dropped before it saturated
- Last model added
- February 2026
In the authors’ words
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.
GPT-5.3 Codex leads with 81.4% among models released by February 2026
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
The newest model with a score here came out in February 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6 Astra
- Claude Opus 5.5
- and 16 more
- 1 GPT-5.3 Codex 81.4%
- 2 GPT-5.1-Codex-Max 80%
- 3 GPT-5 55%
- 4 o3 54%
- 5 OpenAI deep research 53%
- 6 o4-mini 52%
- 7 GPT-5 mini 51%
- 8
Seed2.0 Pro 49.4% - 9 Kimi K2 Jul 2025 39.1%
- 10 GPT-4.5 32.6%
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.3 Codex | 81.4% | Feb 2026 | Lab model card | |
| 2 | GPT-5.1-Codex-Max | 80% | Nov 2025 | Lab model card | |
| 3 | GPT-5 | 55% | Aug 2025 | Lab model card | |
| 4 | o3 | 54% | Apr 2025 | Lab model card | |
| 5 | OpenAI deep research | 53% | Feb 2025 | Lab model card | |
| 6 | o4-mini | 52% | Apr 2025 | Lab model card | |
| 7 | GPT-5 mini | 51% | Aug 2025 | Lab model card | |
| 8 | Seed2.0 Pro | 49.4% | Feb 2026 | Lab model card | |
| 9 | Kimi K2 Jul 2025 | 39.1% | Jul 2025 | Lab model card | |
| 10 | GPT-4.5 | 32.6% | Feb 2025 | Lab model card |
5 models have held number one in labs’ own numbers since its release in February 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Labs’ model cards
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Labs’ model cards
Labs headline it less often now, last in February 2026
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| Feb 14, 2026 | Seed2.0 Pro | 49.4 | SWE-Lancer |
| Feb 5, 2026 | GPT-5.3 Codex | 81.4 | SWE-Lancer IC Diamond |
| Nov 19, 2025 | GPT-5.1-Codex-Max | 79.9 | SWE-Lancer IC SWE, xhigh |
| Aug 7, 2025 | GPT-5 nano | SWE-Lancer: IC SWE Diamond Freelance Coding Tasks ($49K) | |
| Aug 7, 2025 | GPT-5 mini | SWE-Lancer: IC SWE Diamond Freelance Coding Tasks ($75K) | |
| Apr 16, 2025 | o4-mini | SWE-Lancer: IC SWE Diamond (o4-mini-high), dollars earned | |
| Apr 16, 2025 | o3 | SWE-Lancer: IC SWE Diamond (dollars earned), o3-high | |
| Apr 14, 2025 | GPT-4.1 nano | 3.7 | SWE-Lancer (IC-Diamond subset) $9K (3.7%) |
| Apr 14, 2025 | GPT-4.1 mini | 13.1 | SWE-Lancer (IC-Diamond subset) $31K (13.1%) |
| Feb 27, 2025 | GPT-4.5 | 32.6 | SWE-Lancer Diamond (coding)* |
9 system cards report it, from 3 labs
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Feb 14, 2026 | Seed2.0 Pro model card | 49.4 | SWE-Lancer (Table 11) | Capabilities |
| Nov 18, 2025 | GPT-5.1-Codex-Max system card | 80 | SWE-Lancer IC SWE Diamond, pass@1 avg of 3, printed chart label (Figure 8, 5.1.3.1) | AI R&D risk |
| Aug 7, 2025 | GPT-5 system card | 55 | SWE-Lancer IC SWE Diamond, pass@1 avg of 3, no browsing launch candidate (Figure 24) | AI R&D risk |
| Aug 7, 2025 | GPT-5 mini system card | 51 | SWE-Lancer IC SWE Diamond, pass@1 avg of 3, gpt-5-thinking-mini launch candidate (Figure 24) | AI R&D risk |
| Jul 21, 2025 | Kimi K2 technical report | 39.1 | SWE-Lancer (Pass@1) (Table 3) | Capabilities |
| Apr 16, 2025 | o4-mini system card | 52 | SWE-Lancer Diamond (IC SWE), pass@1, o4-mini no browsing launch candidate, printed chart label (Figure 21) | AI R&D risk |
| Apr 16, 2025 | o3 system card | 54 | SWE-Lancer Diamond (IC SWE), pass@1, o3 no browsing launch candidate, printed chart label (Figure 21) | AI R&D risk |
| Feb 27, 2025 | GPT-4.5 system card | 20 | SWE-Lancer Diamond IC SWE tasks, pass@1, post-mitigation, earned $41,625 (4.6.6) | Autonomy |
| Feb 25, 2025 | OpenAI deep research system card | 53 | SWE-Lancer Diamond (IC SWE), pass@1 averaged over 3 runs, post-mitigation with browsing, $87,708 earned (without browsing 48%, $75,583), July 2025 dataset update (3.4.7) | Autonomy |
More in Coding
- SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
- LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests.
DeepSeek-V4-Pro 93.5% - SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
- DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
- SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
- Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471