Code Interpreter Benchmark
Qwen-14B 81.7% leads since Sep 2023 Dropped Models use a Python code interpreter for math, data visualization, and general tasks. Scoring covers result accuracy and the share of code that runs.
The visualization judge changed from Qwen-VL-Chat to GPT-4V in December 2023, so versions differ.
- Level
- School Students, from grade school to high school
- 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.
Qwen-14B leads with 81.7% among models released by Feb 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 Feb 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 Qwen-14B 81.7%
- 2 Qwen1.5-72B 73.1%
- 3 Qwen-7B 70.2%
- 4 Qwen-1_8B 26.8%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen-14B | 81.7% | Sep 2023 | Model card | |
| 2 | Qwen1.5-72B | 73.1% | Feb 2024 | Model card | |
| 3 | Qwen-7B | 70.2% | Sep 2023 | Model card | |
| 4 | Qwen-1_8B | 26.8% | Sep 2023 | Model card |
Number one over time
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.
No launch post has listed it since Feb 2024
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 |
|---|---|---|---|
| 5 Feb 2024 | Qwen1.5-72B | 73.1 | Code Interpreter Benchmark (Accuracy of Code Execution Results, Math) |
| 25 Sep 2023 | Qwen-14B | 58.4 | Code Interpreter benchmark, Accuracy of Code Execution Results (Math) |
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 |
|---|---|---|---|---|
| 28 Sep 2023 | Qwen-7B technical report | 70.2 | In-house Code Interpreter benchmark, executable code rate All (Math 82.4, Visualization 64.4, General 67.2) (Table 7). Final-response correctness: Math 41.9, Vis.-Hard 40.5, Vis.-Easy 54.4, Vis.-All 47.2 (Table 8) | Capabilities |
| 28 Sep 2023 | Qwen-14B technical report | 81.7 | In-house Code Interpreter benchmark, executable code rate All (Math 89.2, Visualization 84.1, General 65.5) (Table 7). Final-response correctness: Math 58.4, Vis.-Hard 53.6, Vis.-Easy 59.5, Vis.-All 56.4 (Table 8) | Capabilities |
| 28 Sep 2023 | Qwen-1_8B technical report | 26.8 | In-house Code Interpreter benchmark, executable code rate All (Math 33.8, Visualization 30.1, General 8.6) (Table 7). Final-response correctness: Math 14.7, Vis.-Hard 3.6, Vis.-Easy 20.3, Vis.-All 11.7 (Table 8) | Capabilities |
More in Tool use
- AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state.
Gemini 4 Carbon 53.9% - MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks.
Muse Spark 1.1 88.1% - BFCL V4 In use Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes. Qwen3.7-Max 75%
- τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state. Qwen 3.8 Max 55.1%
- Claw-Eval In use 300 human-verified agent tasks (service orchestration, multimodal, multi-turn dialogue) audited via traces and snapshots. Rubric scores with Pass^3 consistency. Claude Sonnet 4.6 81.4%
- DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%