llmboard.aiAI model intelligence
Home

Model Rankings

OverallOpen ModelsAgentCodingReasoningMathKnowledgeInstruction FollowingTextVision
Image GenerationImage Editing
Video GenerationImage to VideoVideo Editing
Text to SpeechSpeech to Text
Embeddings

Efficiency

Chat Token PricingImage PricingVideo PricingAudio Pricing
Chat Speed & LatencyProvider Reliability

Benchmarks

GPQAMMLU-ProAIME 2025SWE-Bench VerifiedMMLUHumanity's Last ExamLiveCodeBenchMATHHumanEvalMMMU-Pro
All Benchmarks

Tools

Model Directory

Scoring & Data

Scoring & Data
1224 models729 benchmarks

Leaderboard Center

Overall RankingCodingCore BenchmarksPrice & ValueRuntime Performance

Modalities

All ModelsImage GenerationImage EditingVideo GenerationImage-to-VideoVideo EditingText-to-SpeechSpeech-to-TextEmbeddings

Data & Methods

Scoring MethodAll Benchmarks
llmboard.aiCopyright 2026 llmboard.ai

semantic similarity benchmark

MTEB English v2 Semantic Similarity Leaderboard

MTEB English v2 Semantic Similarity is the current MTEB(eng, v2) semantic_similarity leaderboard.

Updated Sep 8, 2026

Models100
Model coverage211
MetricOfficial leaderboard task-type mean
EvidenceA

On this page

  • Ranking
  • Highlights
  • Distribution
  • Top models
  • About
  • FAQ

MTEB English v2 Semantic Similarity Ranking

Higher official leaderboard task-type mean ranks better on this benchmark.

30 of 100 rows
Columns

Show columns

Sort by
Rank
Model
Official leaderboard task-type mean
Percentile
Participants
Evidence
Evaluated
Rank01ModelKLQZhou-EmbeddingKingsoft LlmOfficial leaderboard task-type mean91.65%Percentile100.00%Participants211EvidenceAEvaluatedN/A
Rank02ModelJCingot-8b-r3JcornersOfficial leaderboard task-type mean89.32%Percentile99.52%Participants211EvidenceAEvaluatedN/A
Rank03ModelINJasper-Token-Compression-600MInfgradOfficial leaderboard task-type mean88.79%Percentile99.05%Participants211EvidenceAEvaluatedN/A
Rank04ModelACQwen3-Embedding-4BAlibaba Cloud / Qwen TeamOfficial leaderboard task-type mean88.72%Percentile98.57%Participants211EvidenceAEvaluatedN/A
Rank05ModelACQwen3-Embedding-8BAlibaba Cloud / Qwen TeamOfficial leaderboard task-type mean88.58%Percentile98.10%Participants211EvidenceAEvaluatedN/A
Rank06ModelJIjina-embeddings-v5-omni-nanoJinaaiOfficial leaderboard task-type mean88.28%Percentile97.62%Participants211EvidenceAEvaluatedN/A
Rank07ModelJIjina-embeddings-v5-text-nanoJinaaiOfficial leaderboard task-type mean88.28%Percentile97.14%Participants211EvidenceAEvaluatedN/A
Rank08ModelJIjina-embeddings-v5-omni-smallJinaaiOfficial leaderboard task-type mean88.13%Percentile96.67%Participants211EvidenceAEvaluatedN/A
Rank09ModelJIjina-embeddings-v5-text-smallJinaaiOfficial leaderboard task-type mean88.13%Percentile96.19%Participants211EvidenceAEvaluatedN/A
Rank10ModelBSSeed1.5-EmbeddingBytedance SeedOfficial leaderboard task-type mean87.23%Percentile95.71%Participants211EvidenceAEvaluatedN/A
Rank11ModelBYSeed1.6-embeddingByteDanceOfficial leaderboard task-type mean86.87%Percentile95.24%Participants211EvidenceAEvaluatedN/A
Rank12ModelANLGAI-Embedding-PreviewAnnamodelsOfficial leaderboard task-type mean86.69%Percentile94.76%Participants211EvidenceAEvaluatedN/A
Rank13ModelACQwen3-Embedding-0.6BAlibaba Cloud / Qwen TeamOfficial leaderboard task-type mean86.57%Percentile94.29%Participants211EvidenceAEvaluatedN/A
Rank14ModelLAbilingual-embedding-largeLajavanessOfficial leaderboard task-type mean86.00%Percentile93.81%Participants211EvidenceAEvaluatedN/A
Rank15ModelJIjina-embeddings-v4JinaaiOfficial leaderboard task-type mean85.89%Percentile93.33%Participants211EvidenceAEvaluatedN/A
Rank16ModelJIjina-embeddings-v3JinaaiOfficial leaderboard task-type mean85.82%Percentile92.86%Participants211EvidenceAEvaluatedN/A
Rank17ModelKINano-Em1-0.6B-v2.1KitefishaiOfficial leaderboard task-type mean85.44%Percentile92.38%Participants211EvidenceAEvaluatedN/A
Rank18ModelGOGemini Embedding 001GoogleOfficial leaderboard task-type mean85.29%Percentile91.90%Participants211EvidenceAEvaluatedN/A
Rank19ModelGOtext-embedding-005GoogleOfficial leaderboard task-type mean85.18%Percentile91.43%Participants211EvidenceAEvaluatedN/A
Rank20ModelIEYuan-embedding-2.0-enIeityuanOfficial leaderboard task-type mean84.89%Percentile90.95%Participants211EvidenceAEvaluatedN/A
Rank21ModelGOtext-embedding-004GoogleOfficial leaderboard task-type mean84.84%Percentile90.48%Participants211EvidenceAEvaluatedN/A
Rank22ModelASGiga-Embeddings-instruct-3B-0826Ai SageOfficial leaderboard task-type mean84.83%Percentile90.00%Participants211EvidenceAEvaluatedN/A
Rank23ModelKEKaLM-embedding-multilingual-mini-instruct-v2.5Kalm EmbeddingOfficial leaderboard task-type mean84.82%Percentile89.52%Participants211EvidenceAEvaluatedN/A
Rank24ModelSASFR-Embedding-MistralSalesforceOfficial leaderboard task-type mean84.77%Percentile89.05%Participants211EvidenceAEvaluatedN/A
Rank25ModelHCspeed-embedding-7b-instructHaon ChenOfficial leaderboard task-type mean84.74%Percentile88.57%Participants211EvidenceAEvaluatedN/A
Rank26ModelINmultilingual-e5-large-instructIntfloatOfficial leaderboard task-type mean84.72%Percentile88.10%Participants211EvidenceAEvaluatedN/A
Rank27ModelLALinq-Embed-MistralLinq Ai ResearchOfficial leaderboard task-type mean84.69%Percentile87.62%Participants211EvidenceAEvaluatedN/A
Rank28ModelCAF2LLM-v2-8BCodefuse AiOfficial leaderboard task-type mean84.65%Percentile87.14%Participants211EvidenceAEvaluatedN/A
Rank29ModelBIBidirLM-Omni-2.5B-EmbeddingBidirlmOfficial leaderboard task-type mean84.62%Percentile86.67%Participants211EvidenceAEvaluatedN/A
Rank30ModelTATarka-Embedding-350M-V1Tarka AirOfficial leaderboard task-type mean84.59%Percentile86.19%Participants211EvidenceAEvaluatedN/A

MTEB English v2 Semantic Similarity Highlights

The leading models and scores on this benchmark.

MTEB English v2 Semantic Similarity Score Distribution

A closer view of the leading scores on this benchmark.

MTEB English v2 Semantic Similarity

The Top AI Models for MTEB English v2 Semantic Similarity

The first five results on this benchmark, with official price and output speed added where the model identity can be matched.

What is MTEB English v2 Semantic Similarity?

What MTEB English v2 Semantic Similarity measures and how its scores work.

MTEB English v2 Semantic Similarity is a semantic_similarity benchmark from MTEB(eng, v2).

It reports the Official leaderboard task-type mean as a ratio.

Scores are shown in ratio. This benchmark is verified and has an evidence level of A.

Family
MTEB English v2
Modality
text
Primary category
semantic similarity
Score direction
higher
LLMBoard eligible
No
Evaluation key
overall

MTEB. Benchmark scores retain their original unit. Overall score eligibility is shown separately.

FAQ

Common questions about MTEB English v2 Semantic Similarity.

Which model scores highest on MTEB English v2 Semantic Similarity?

QZhou-Embedding is currently ranked first with 91.65%.

What are the top three models on MTEB English v2 Semantic Similarity?

The current leaders are QZhou-Embedding (91.65%), ingot-8b-r3 (89.32%), and Jasper-Token-Compression-600M (88.79%).

Which MTEB English v2 Semantic Similarity model has the lowest official input price?

Gemini Embedding 001 has the lowest matched official input price at $0.15 input per 1M tokens.

Which models are fastest among MTEB English v2 Semantic Similarity results?

No matched runtime record is currently available.

Does the highest official leaderboard task-type mean result prove overall model quality?

No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.

What does MTEB English v2 Semantic Similarity measure?

It reports the Official leaderboard task-type mean as a ratio.

Is a higher official leaderboard task-type mean better?

Yes. Higher values rank better for this benchmark.

How many models are compared?

100 model results are currently shown.

Does this benchmark affect the overall score?

No. This benchmark is shown for reference but does not contribute to the overall score.

Rank #1QZhou-Embedding91.65%
Rank #2ingot-8b-r389.32%
Rank #3Jasper-Token-Compression-600M88.79%
Rank #4Qwen3-Embedding-4B88.72%

Ranking basisThis mteb english v2 semantic similarity AI model leaderboard uses descending official leaderboard task-type mean in the benchmark original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.

  1. 01
    KL
    QZhou-EmbeddingKingsoft Llm
    Official leaderboard task-type mean
    91.65%

    Strengths

    • Ranks #1 of 211 compared models
    • 100th percentile on this benchmark
    • A evidence result

    Considerations

    • This result measures MTEB English v2 Semantic Similarity, not total model capability
  2. 02
    JC
    ingot-8b-r3Jcorners
    Official leaderboard task-type mean
    89.32%

    Strengths

    • Ranks #2 of 211 compared models
    • 100th percentile on this benchmark
    • A evidence result

    Considerations

    • This result measures MTEB English v2 Semantic Similarity, not total model capability
  3. 03
    IN
    Infgrad
    Official leaderboard task-type mean
    88.79%

    Strengths

    • Ranks #3 of 211 compared models
    • 99th percentile on this benchmark
    • A evidence result

    Considerations

    • This result measures MTEB English v2 Semantic Similarity, not total model capability
  4. 04
    AC
    Alibaba Cloud / Qwen Team
    Official leaderboard task-type mean
    88.72%

    Strengths

    • Ranks #4 of 211 compared models
    • 99th percentile on this benchmark
    • A evidence result

    Considerations

    • This result measures MTEB English v2 Semantic Similarity, not total model capability
  5. 05
    AC
    Alibaba Cloud / Qwen Team
    Official leaderboard task-type mean
    88.58%

    Strengths

    • Ranks #5 of 211 compared models
    • 98th percentile on this benchmark
    • A evidence result

    Considerations

    • This result measures MTEB English v2 Semantic Similarity, not total model capability

Selection summary

Best AI Models for MTEB English v2 Semantic Similarity

QZhou-Embedding currently leads MTEB English v2 Semantic Similarity with 91.65%. 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.

Jasper-Token-Compression-600M
Qwen3-Embedding-4B
Qwen3-Embedding-8B
Benchmark rank #1QZhou-Embedding91.65%
Benchmark rank #2ingot-8b-r389.32%
Benchmark rank #3Jasper-Token-Compression-600M88.79%