math benchmark
GSM8k Leaderboard
Grade School Math 8K, a dataset of 8.5K high-quality linguistically diverse grade school math word problems requiring multi-step reasoning and elementary arithmetic operations.
Updated Aug 17, 2026
Models48
Model coverage48
MetricScore
EvidenceB
GSM8k Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelXI | Score99.6% | Percentile100.0% | Participants48 | EvidenceC | Evaluated |
| Rank02 | ModelMA | Score97.3% | Percentile97.9% | Participants48 | EvidenceC | Evaluated |
| Rank03 | ModelOP | Score97.1% | Percentile95.7% | Participants48 | EvidenceC | Evaluated |
| Rank04 | ModelOP | Score97.0% | Percentile93.6% | Participants48 | EvidenceC | Evaluated |
| Rank05 | ModelME | Score96.8% | Percentile91.5% | Participants48 | EvidenceC | Evaluated |
| Rank06 | ModelAN | Score96.4% | Percentile89.4% | Participants48 | EvidenceC | Evaluated |
| Rank07 | ModelAN | Score96.4% | Percentile87.2% | Participants48 | EvidenceC | Evaluated |
| Rank08 | ModelGO | Score95.9% | Percentile85.1% | Participants48 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score95.9% | Percentile83.0% | Participants48 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Score95.8% | Percentile80.8% | Participants48 | EvidenceC | Evaluated |
| Rank11 | ModelDE | Score95.1% | Percentile78.7% | Participants48 | EvidenceC | Evaluated |
| Rank12 | ModelAN | Score95.0% | Percentile76.6% | Participants48 | EvidenceC | Evaluated |
| Rank13 | ModelAM Nova ProAmazon | Score94.8% | Percentile74.5% | Participants48 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score94.8% | Percentile72.3% | Participants48 | EvidenceC | Evaluated |
| Rank15 | ModelAMNova LiteAmazon | Score94.5% | Percentile70.2% | Participants48 | EvidenceC | Evaluated |
| Rank16 | ModelGO | Score94.4% | Percentile68.1% | Participants48 | EvidenceC | Evaluated |
| Rank17 | ModelAC | Score94.4% | Percentile66.0% | Participants48 | EvidenceC | Evaluated |
| Rank18 | ModelMA | Score93.0% | Percentile63.8% | Participants48 | EvidenceC | Evaluated |
| Rank19 | ModelAN | Score92.3% | Percentile61.7% | Participants48 | EvidenceC | Evaluated |
| Rank20 | ModelAMNova MicroAmazon | Score92.3% | Percentile59.6% | Participants48 | EvidenceC | Evaluated |
| Rank21 | ModelMA | Score92.1% | Percentile57.5% | Participants48 | EvidenceC | Evaluated |
| Rank22 | ModelAC | Score91.6% | Percentile55.3% | Participants48 | EvidenceC | Evaluated |
| Rank23 | ModelNV | Score91.4% | Percentile53.2% | Participants48 | EvidenceC | Evaluated |
| Rank24 | ModelAC | Score91.1% | Percentile51.1% | Participants48 | EvidenceC | Evaluated |
| Rank25 | ModelAC | Score91.1% | Percentile48.9% | Participants48 | EvidenceC | Evaluated |
| Rank26 | ModelGO | Score90.8% | Percentile46.8% | Participants48 | EvidenceC | Evaluated |
| Rank27 | ModelXA | Score90.0% | Percentile44.7% | Participants48 | EvidenceC | Evaluated |
| Rank28 | ModelGO | Score89.2% | Percentile42.5% | Participants48 | EvidenceC | Evaluated |
| Rank29 | ModelAN | Score88.9% | Percentile40.4% | Participants48 | EvidenceC | Evaluated |
| Rank30 | ModelMI | Score88.7% | Percentile38.3% | Participants48 | EvidenceC | Evaluated |
| Rank31 | ModelAC | Score88.7% | Percentile36.2% | Participants48 | EvidenceC | Evaluated |
| Rank32 | ModelMI | Score88.6% | Percentile34.0% | Participants48 | EvidenceC | Evaluated |
| Rank33 | ModelAL | Score87.0% | Percentile31.9% | Participants48 | EvidenceC | Evaluated |
| Rank34 | ModelGO | Score86.2% | Percentile29.8% | Participants48 | EvidenceC | Evaluated |
| Rank35 | ModelMI | Score86.2% | Percentile27.7% | Participants48 | EvidenceC | Evaluated |
| Rank36 | ModelAC | Score83.9% | Percentile25.5% | Participants48 | EvidenceC | Evaluated |
| Rank37 | ModelAC | Score82.3% | Percentile23.4% | Participants48 | EvidenceC | Evaluated |
| Rank38 | ModelIB | Score80.9% | Percentile21.3% | Participants48 | EvidenceC | Evaluated |
| Rank39 | ModelMA | Score80.7% | Percentile19.1% | Participants48 | EvidenceC | Evaluated |
| Rank40 | ModelME | Score77.7% | Percentile17.0% | Participants48 | EvidenceC | Evaluated |
| Rank41 | ModelAL | Score75.8% | Percentile14.9% | Participants48 | EvidenceC | Evaluated |
| Rank42 | ModelGO | Score74.0% | Percentile12.8% | Participants48 | EvidenceC | Evaluated |
| Rank43 | ModelCO | Score70.7% | Percentile10.6% | Participants48 | EvidenceC | Evaluated |
| Rank44 | ModelIB | Score70.1% | Percentile8.5% | Participants48 | EvidenceC | Evaluated |
| Rank45 | ModelGO | Score68.6% | Percentile6.4% | Participants48 | EvidenceC | Evaluated |
| Rank46 | ModelGO | Score62.8% | Percentile4.3% | Participants48 | EvidenceC | Evaluated |
| Rank47 | ModelIB | Score59.0% | Percentile2.1% | Participants48 | EvidenceC | Evaluated |
| Rank48 | ModelBA | Score25.2% | Percentile0.0% | Participants48 | EvidenceC | Evaluated |
GSM8k Highlights
The leading models and scores on this benchmark.
GSM8k Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for GSM8k
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis gsm8k AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01XIMiMo-V2.5-ProXiaomi
- Score
- 99.6%
- Price
- $0.43 input / $0.87 output per 1M tokens
- Speed
- Up to 69 tok/s via Novita
Strengths
- Ranks #1 of 48 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures GSM8k, not total model capability
- 02MAKimi K2 InstructMoonshot AI
- Score
- 97.3%
- Price
- $0.60 input / $2.5 output per 1M tokens
- Speed
- Up to 45 tok/s via Novita
Strengths
- Ranks #2 of 48 compared models
- 98th percentile on this benchmark
- C evidence result
Considerations
- This result measures GSM8k, not total model capability
- 03OPo1OpenAI
- Score
- 97.1%
- Price
- $15 input / $60 output per 1M tokens
- Speed
- Up to 66 tok/s via OpenAI
Strengths
- Ranks #3 of 48 compared models
- 96th percentile on this benchmark
- C evidence result
Considerations
- This result measures GSM8k, not total model capability
- 04OPGPT-4.5OpenAI
- Score
- 97.0%
- Speed
- Up to 50 tok/s via OpenAI
Strengths
- Ranks #4 of 48 compared models
- 94th percentile on this benchmark
- C evidence result
Considerations
- This result measures GSM8k, not total model capability
- 05ME
- Score
- 96.8%
- Speed
- Up to 100 tok/s via Bedrock
Strengths
- Ranks #5 of 48 compared models
- 91th percentile on this benchmark
- C evidence result
Considerations
- This result measures GSM8k, not total model capability
Selection summary
Best AI Models for GSM8k
MiMo-V2.5-Pro currently leads GSM8k with 99.6%. 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 GSM8k?
What GSM8k measures and how its scores work.
Grade School Math 8K, a dataset of 8.5K high-quality linguistically diverse grade school math word problems requiring multi-step reasoning and elementary arithmetic operations.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- GSM8k
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- gsm8k|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about GSM8k.
Which model scores highest on GSM8k?
MiMo-V2.5-Pro is currently ranked first with 99.6%.
What does GSM8k measure?
Grade School Math 8K, a dataset of 8.5K high-quality linguistically diverse grade school math word problems requiring multi-step reasoning and elementary arithmetic operations.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
48 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.