reasoning benchmark
MBPP Leaderboard
MBPP (Mostly Basic Python Problems) is a benchmark of 974 crowd-sourced Python programming problems designed to be solvable by entry-level programmers. Each problem consists of a task description, code solution, and 3 automated test cases covering programming fundamentals and standard library functionality.
Updated Aug 17, 2026
Models33
Model coverage33
MetricPass@1
EvidenceB
MBPP Ranking
Higher pass@1 ranks better on this benchmark.
Sort by
Rank | Model | Pass@1 | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelSA | Pass@192.7% | Percentile100.0% | Participants33 | EvidenceC | Evaluated |
| Rank02 | ModelNV | Pass@191.3% | Percentile96.9% | Participants33 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Pass@190.2% | Percentile93.8% | Participants33 | EvidenceC | Evaluated |
| Rank04 | ModelOP | Pass@189.1% | Percentile90.6% | Participants33 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Pass@188.2% | Percentile87.5% | Participants33 | EvidenceC | Evaluated |
| Rank06 | ModelNV | Pass@184.6% | Percentile84.4% | Participants33 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Pass@184.0% | Percentile81.3% | Participants33 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Pass@184.0% | Percentile78.1% | Participants33 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Pass@183.5% | Percentile75.0% | Participants33 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Pass@182.0% | Percentile71.9% | Participants33 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Pass@181.4% | Percentile68.8% | Participants33 | EvidenceC | Evaluated |
| Rank12 | ModelMI | Pass@180.8% | Percentile65.6% | Participants33 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Pass@180.2% | Percentile62.5% | Participants33 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Pass@179.2% | Percentile59.4% | Participants33 | EvidenceC | Evaluated |
| Rank15 | ModelMA | Pass@178.2% | Percentile56.3% | Participants33 | EvidenceC | Evaluated |
| Rank16 | ModelME | Pass@177.6% | Percentile53.1% | Participants33 | EvidenceC | Evaluated |
| Rank17 | ModelGO | Pass@176.0% | Percentile50.0% | Participants33 | EvidenceC | Evaluated |
| Rank18 | ModelMA | Pass@174.7% | Percentile46.9% | Participants33 | EvidenceC | Evaluated |
| Rank19 | ModelGO | Pass@174.4% | Percentile43.8% | Participants33 | EvidenceC | Evaluated |
| Rank20 | ModelAC | Pass@173.2% | Percentile40.6% | Participants33 | EvidenceC | Evaluated |
| Rank21 | ModelGO | Pass@173.0% | Percentile37.5% | Participants33 | EvidenceC | Evaluated |
| Rank22 | ModelMA | Pass@169.6% | Percentile34.4% | Participants33 | EvidenceC | Evaluated |
| Rank23 | ModelMI | Pass@169.6% | Percentile31.3% | Participants33 | EvidenceC | Evaluated |
| Rank24 | ModelME | Pass@167.8% | Percentile28.1% | Participants33 | EvidenceC | Evaluated |
| Rank25 | ModelAC | Pass@167.2% | Percentile25.0% | Participants33 | EvidenceC | Evaluated |
| Rank26 | ModelGO | Pass@163.6% | Percentile21.9% | Participants33 | EvidenceC | Evaluated |
| Rank27 | ModelGO | Pass@163.6% | Percentile18.8% | Participants33 | EvidenceC | Evaluated |
| Rank28 | ModelGO | Pass@163.2% | Percentile15.6% | Participants33 | EvidenceC | Evaluated |
| Rank29 | ModelGO | Pass@162.6% | Percentile12.5% | Participants33 | EvidenceC | Evaluated |
| Rank30 | ModelGO | Pass@156.6% | Percentile9.4% | Participants33 | EvidenceC | Evaluated |
| Rank31 | ModelGO | Pass@156.6% | Percentile6.3% | Participants33 | EvidenceC | Evaluated |
| Rank32 | ModelGO | Pass@152.4% | Percentile3.1% | Participants33 | EvidenceC | Evaluated |
| Rank33 | ModelGO | Pass@135.2% | Percentile0.0% | Participants33 | EvidenceC | Evaluated |
MBPP Highlights
The leading models and scores on this benchmark.
MBPP Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for MBPP
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis mbpp AI model leaderboard uses descending pass@1 in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01SASarvam-30BSarvam AI
- Pass@1
- 92.7%
Strengths
- Ranks #1 of 33 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures MBPP, not total model capability
- 02NV
- Pass@1
- 91.3%
Strengths
- Ranks #2 of 33 compared models
- 97th percentile on this benchmark
- C evidence result
Considerations
- This result measures MBPP, not total model capability
- 03ACQwen2.5-Coder 32B InstructAlibaba Cloud / Qwen Team
- Pass@1
- 90.2%
- Speed
- Up to 110 tok/s via Fireworks
Strengths
- Ranks #3 of 33 compared models
- 94th percentile on this benchmark
- C evidence result
Considerations
- This result measures MBPP, not total model capability
- 04OPMiniCPM-SALAOpenBMB
- Pass@1
- 89.1%
Strengths
- Ranks #4 of 33 compared models
- 91th percentile on this benchmark
- C evidence result
Considerations
- This result measures MBPP, not total model capability
- 05ACQwen2.5 72B InstructAlibaba Cloud / Qwen Team
- Pass@1
- 88.2%
- Price
- $1.4 input / $5.6 output per 1M tokens
- Speed
- Up to 100 tok/s via Hyperbolic
Strengths
- Ranks #5 of 33 compared models
- 88th percentile on this benchmark
- C evidence result
Considerations
- This result measures MBPP, not total model capability
Selection summary
Best AI Models for MBPP
Sarvam-30B currently leads MBPP with 92.7%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is MBPP?
What MBPP measures and how its scores work.
MBPP (Mostly Basic Python Problems) is a benchmark of 974 crowd-sourced Python programming problems designed to be solvable by entry-level programmers. Each problem consists of a task description, code solution, and 3 automated test cases covering programming fundamentals and standard library functionality.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MBPP
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- mbpp|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MBPP.
Which model scores highest on MBPP?
Sarvam-30B is currently ranked first with 92.7%.
What does MBPP measure?
MBPP (Mostly Basic Python Problems) is a benchmark of 974 crowd-sourced Python programming problems designed to be solvable by entry-level programmers. Each problem consists of a task description, code solution, and 3 automated test cases covering programming fundamentals and standard library functionality.
Is a higher pass@1 better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
33 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.