Back to all embeddings

Kanon 2

Isaacus Kanon 2 embedding model optimized for legal and regulatory text retrieval. If you want to compare the best embedding models for your data, try Agentset.

Leaderboard Rank
#14
of 17
ELO Rating
1450
#14
Win Rate
33.5%
#16
Accuracy (nDCG@10)
0.484
#17
Latency
250ms
#14

Model Information

Provider
Isaacus
License
Open Source
Price per 1M tokens
$0.350
Dimensions
1792
Release Date
2025-10-16
Model Name
kanon-2
Total Evaluations
1139

Performance Record

Wins381 (33.5%)
Losses542 (47.6%)
Ties216 (19.0%)
Wins
Losses
Ties

Embedding Models Are Just One Piece of RAG

Agentset gives you a managed RAG pipeline with the top-ranked models and best practices baked in. No infrastructure to maintain, no embeddings to manage.

Trusted by teams building production RAG applications

5M+
Documents
1,500+
Teams
99.9%
Uptime

Performance Overview

ELO ratings by dataset

Kanon 2's ELO performance varies across different benchmark datasets, showing its strengths in specific domains.

Kanon 2 - ELO by Dataset

Detailed Metrics

Dataset breakdown

Performance metrics across different benchmark datasets, including accuracy and latency percentiles.

PG

ELO 150033.3% WR20W-20L-20T

Accuracy Metrics

nDCG@5
0.000
nDCG@10
0.000
Recall@5
0.000
Recall@10
0.000

Latency Distribution

Mean
250ms
P50 (Median)
250ms
P90
250ms

business reports

ELO 150050.0% WR90W-57L-33T

Accuracy Metrics

nDCG@5
0.000
nDCG@10
0.000
Recall@5
0.000
Recall@10
0.000

Latency Distribution

Mean
250ms
P50 (Median)
250ms
P90
250ms

DBPedia

ELO 150033.3% WR60W-77L-43T

Accuracy Metrics

nDCG@5
0.806
nDCG@10
0.777
Recall@5
0.062
Recall@10
0.120

Latency Distribution

Mean
250ms
P50 (Median)
250ms
P90
250ms

FiQa

ELO 150056.7% WR102W-48L-30T

Accuracy Metrics

nDCG@5
0.839
nDCG@10
0.836
Recall@5
0.689
Recall@10
0.763

Latency Distribution

Mean
250ms
P50 (Median)
250ms
P90
250ms

SciFact

ELO 150032.8% WR59W-93L-28T

Accuracy Metrics

nDCG@5
0.718
nDCG@10
0.744
Recall@5
0.772
Recall@10
0.861

Latency Distribution

Mean
250ms
P50 (Median)
250ms
P90
250ms

MSMARCO

ELO 150027.9% WR50W-98L-31T

Accuracy Metrics

nDCG@5
0.941
nDCG@10
0.931
Recall@5
0.117
Recall@10
0.223

Latency Distribution

Mean
250ms
P50 (Median)
250ms
P90
250ms

ARCD

ELO 15000.0% WR0W-149L-31T

Accuracy Metrics

nDCG@5
0.009
nDCG@10
0.009
Recall@5
0.020
Recall@10
0.020

Latency Distribution

Mean
250ms
P50 (Median)
250ms
P90
250ms

Build RAG in Minutes, Not Months

Agentset gives you a complete RAG API with top-ranked embedding models and smart retrieval built in. Upload your data, call the API, and get accurate results from day one.

import { Agentset } from "agentset";

const agentset = new Agentset();
const ns = agentset.namespace("ns_1234");

const results = await ns.search(
  "What is multi-head attention?"
);

for (const result of results) {
  console.log(result.text);
}

Compare Models

See how it stacks up

Compare Kanon 2 with other top embeddings to understand the differences in performance, accuracy, and latency.