tensor.news
Integrity IndexModelsCompareLabsPapersMethodologyAbout
tensor.news

© 2026

Evidence

  • Integrity Index
  • Model Claim Ledger
  • Compare

Entities

  • Models
  • Benchmarks
  • Labs
  • Papers

Method

  • Methodology
  • About
  • Privacy

Machine

  • Feed (Atom)
  • llms.txt
  • llms-full.txt
  • Sitemap

tensor.news grades benchmarks and tags model claims by evidence status. Every fact carries a source record and provenance. Scores are framed as task performance under a disclosed harness, never as deployed capability.

← Integrity Index

CL-bench

CL-bench

Rank #29 of 51 · 19 models scored · top score 27.9

BINTEGRITY 82 / 100

unknown

Strongest on contamination resistance, weakest on discrimination. Caveat: the public test set leaves contamination on the table.

consistent harnesspublic test set

Headroom to ceiling

72.1 headroomtop 27.9ceiling 100

Score distribution

median 18.7GPT-5.4 27.91019.529

State of the art over time

10202927.920242026

Integrity breakdown

Weighted composite. Each signal is computed from the source-backed leaderboard below.

Discrimination30%41

Does it still separate models?

Saturation headroom30%99

How far from ceiling / clustered at the top?

Contamination resistance15%100

Public vs held-out; training-leak risk.

Harness comparability15%100

Apples-to-apples, or mixed / vendor-optimized?

Freshness10%100

How old is the benchmark?

Leaderboard

#ModelScoreEvidenceMeasured
1GPT-5.427.9unverified2026-03-05
2GPT-5.123.7unverified2025-11-13
3Grok 4.2022.2unverified2026-02-17
4Claude Opus 4.521.1unverified2025-11-24
5Gemini 3.1 Pro20.8unverified2026-02-19
6Claude Opus 4.620.7unverified2026-02-05
7Qwen 3.6 Plus20.3unverified2026-04-01
8Qwen 3.5 Plus (hosted 397B-A17B)19.8unverified2026-02-16
9Kimi K2.519.3unverified2026-02-02
10GLM-518.7unverified2026-02-11
11GPT-5.218.2unverified2025-12-11
12o317.8unverified2024-12-20
13Kimi K2 Thinking17.6unverified2025-11-06
14GLM-4.715.9unverified2025-12-22
15Gemini 3 Pro15.8unverified2025-11-18
16Qwen3-Max14.5unverified2025-09-05
17DeepSeek-V3.2-Exp13.2unverified2025-09-29
18DeepSeek-V3.212.4unverified2025-12-01
19MiniMax-M2.511.4unverified2026-02-12

At a glance

Models scored
19
Top model
GPT-5.4
Top score
27.9
Score spread
16.5
Headroom
72.1

Benchmark context

Task type
unknown
Metric
normalized performance in Epoch CSV, rendered as accuracy (%)
Score ceiling
100
Human baseline
unknown
Test-set privacy
unknown
Construction
unknown
Key assumptions
unknown
Limitations
unknown
Contamination history
unknown
Harness notes
Epoch CSV reports model, model version, performance, optimized flag, score source, benchmark release date, and model date.
Authors
unknown
Released
unknown
License
CC BY 4.0
Source
epoch.ai/data/eci_benchmarks.csv

Provenance

19 source records · scores from epoch_benchmarks. Every fact on this page is evidence-linked.

Follow the record

Integrity neighbors

  • SimpleBenchA · 86
  • Terminal BenchB · 84
  • SWE-Bench verifiedB · 83
  • CL-bench LifeB · 81
  • GPQA diamondB · 81
  • OSWorld 2.0B · 81

Top models here

  • GPT-5.427.9
  • GPT-5.123.7
  • Grok 4.2022.2
  • Claude Opus 4.521.1
  • Gemini 3.1 Pro20.8

Head-to-head

  • GPT-5.4 vs GPT-5.1
  • GPT-5.4 vs Grok 4.20
  • GPT-5.1 vs Grok 4.20