Internal agentic coding evaluation (Anthropic)
Agents implement a pull request in an open source codebase from a natural language description. A problem counts as solved when all hidden tests pass.
In-house Anthropic evaluation based on real pull requests, first reported in the Claude 3.5 Sonnet model card addendum. The tasks are not public.
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
- Answers
- Agent task Many steps in a terminal, browser, desktop, app, game or robot.
Claude 3.5 Sonnet (October 2024) leads with 78% among models released by Oct 2024
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 Oct 2024, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6 Astra
- Claude Opus 5.5
- GPT-6.1 Terra
- GPT-6.1 Sol
- and 16 more
- 1 Claude 3.5 Sonnet (October 2024) 78%
- 2 Claude 3.5 Haiku 74%
- 3 Claude 3.5 Sonnet 64%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude 3.5 Sonnet (October 2024) | 78% | Oct 2024 | Model card | |
| 2 | Claude 3.5 Haiku | 74% | Oct 2024 | Model card | |
| 3 | Claude 3.5 Sonnet | 64% | Jun 2024 | Model card |
2 models have held number one since its release in Jun 2024
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Labs report it only in system cards
We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.
3 system cards report it, from 1 lab
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 |
|---|---|---|---|---|
| 22 Oct 2024 | Claude 3.5 Sonnet (October 2024) model card | 78 | Internal agentic coding evaluation, % of problems which pass all tests (Table 4) | Capabilities |
| 22 Oct 2024 | Claude 3.5 Haiku model card | 74 | Internal agentic coding evaluation, % of problems which pass all tests (Table 4) | Capabilities |
| 20 Jun 2024 | Claude 3.5 Sonnet model card | 64 | Internal agentic coding evaluation, % of problems which pass all tests (Table 3) | Capabilities |
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%
- SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Opus 5.5 93.9%
- 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% - 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