reasoning benchmark
Wild Bench Leaderboard
WildBench is an automated evaluation framework that benchmarks large language models using 1,024 challenging, real-world tasks selected from over one million human-chatbot conversation logs. It introduces two evaluation metrics (WB-Reward and WB-Score) that achieve high correlation with human preferences and uses task-specific checklists for systematic evaluation.
Updated Aug 17, 2026
Models8
Model coverage8
MetricScore
EvidenceB
Wild Bench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelMA | Score68.5% | Percentile100.0% | Participants8 | EvidenceC | Evaluated |
| Rank02 | ModelMA | Score68.5% | Percentile85.7% | Participants8 | EvidenceC | Evaluated |
| Rank03 | ModelMA | Score66.8% | Percentile71.4% | Participants8 | EvidenceC | Evaluated |
| Rank04 | ModelMA | Score65.3% | Percentile57.1% | Participants8 | EvidenceC | Evaluated |
| Rank05 | ModelMA | Score56.8% | Percentile42.9% | Participants8 | EvidenceC | Evaluated |
| Rank06 | ModelMA | Score52.2% | Percentile28.6% | Participants8 | EvidenceC | Evaluated |
| Rank07 | ModelAL | Score48.5% | Percentile14.3% | Participants8 | EvidenceC | Evaluated |
| Rank08 | ModelAL | Score42.4% | Percentile0.0% | Participants8 | EvidenceC | Evaluated |
Wild Bench Highlights
The leading models and scores on this benchmark.
Wild Bench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for Wild Bench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis wild bench 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.
- 01MAMiniStral 3 (14B Instruct 2512)Mistral AI
- Score
- 68.5%
- Price
- $0.10 input / $0.10 output per 1M tokens
- Speed
- Up to 238 tok/s via Mistral AI
Strengths
- Ranks #1 of 8 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures Wild Bench, not total model capability
- 02MAMistral Large 3Mistral AI
- Score
- 68.5%
- Price
- $0.50 input / $1.5 output per 1M tokens
- Speed
- Up to 0.78 tok/s via Mistral AI
Strengths
- Ranks #2 of 8 compared models
- 86th percentile on this benchmark
- C evidence result
Considerations
- This result measures Wild Bench, not total model capability
- 03MAMinistral 3 (8B Instruct 2512)Mistral AI
- Score
- 66.8%
- Price
- $0.10 input / $0.10 output per 1M tokens
- Speed
- Up to 238 tok/s via Mistral AI
Strengths
- Ranks #3 of 8 compared models
- 71th percentile on this benchmark
- C evidence result
Considerations
- This result measures Wild Bench, not total model capability
- 04MAMistral Small 3.2 24B InstructMistral AI
- Score
- 65.3%
Strengths
- Ranks #4 of 8 compared models
- 57th percentile on this benchmark
- C evidence result
Considerations
- This result measures Wild Bench, not total model capability
- 05MAMinistral 3 (3B Instruct 2512)Mistral AI
- Score
- 56.8%
- Price
- $0.10 input / $0.10 output per 1M tokens
- Speed
- Up to 238 tok/s via Mistral AI
Strengths
- Ranks #5 of 8 compared models
- 43th percentile on this benchmark
- C evidence result
Considerations
- This result measures Wild Bench, not total model capability
Selection summary
Best AI Models for Wild Bench
MiniStral 3 (14B Instruct 2512) currently leads Wild Bench with 68.5%. 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 Wild Bench?
What Wild Bench measures and how its scores work.
WildBench is an automated evaluation framework that benchmarks large language models using 1,024 challenging, real-world tasks selected from over one million human-chatbot conversation logs. It introduces two evaluation metrics (WB-Reward and WB-Score) that achieve high correlation with human preferences and uses task-specific checklists for systematic evaluation.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Wild Bench
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- wild-bench|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Wild Bench.
Which model scores highest on Wild Bench?
MiniStral 3 (14B Instruct 2512) is currently ranked first with 68.5%.
What does Wild Bench measure?
WildBench is an automated evaluation framework that benchmarks large language models using 1,024 challenging, real-world tasks selected from over one million human-chatbot conversation logs. It introduces two evaluation metrics (WB-Reward and WB-Score) that achieve high correlation with human preferences and uses task-specific checklists for systematic evaluation.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
8 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.