Benchmark Almanac

Instrumentation Universal Knowledge v1.5

Grok-3 mini 86.4% leads since June 2026
Not in model cards
Version 1.5

Models diagnose instrumentation faults and answer industrial controls questions across four tiers. The composite weights each tier by its difficulty and question count.

RelayForge ranks 15 models on 300 questions from a pool of 1,547. Its October errata preserve the published scores.

Made by
RelayForge
Released
June 2026
Kind
Tasks from real jobs
Level
Expert
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Run by
Anyone
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
15
Entries on its own leaderboard, 14 models we track

Grok-3 mini leads with 86.4% among models released by April 2026 Grok-3 mini leads with 86.4% among models released by April 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in April 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Grok-3 mini xai 86.4%
  2. 2 Claude Opus 4.7 openrouter 85.8%
  3. 3 Claude Sonnet 4.6 openrouter 83.2%
  4. 4 Grok 4.3 openrouter 83.2%
  5. 5 Devstral 2 openrouter 71.4%
  6. 6 Gemma 3 27B openrouter 66%
  7. 7 Mistral Medium 3.5 openrouter 65.9%
  8. 8 GPT-5.5 openrouter 62.9%
  9. 9 Qwen 2.5 32B Instruct ollama_remote 60.1%
  10. 10 Llama 4 Maverick openrouter 60%
  11. 11 Llama 4 Scout openrouter 57.6%
  12. 12 Gemma 3 12B openrouter 55.5%
  13. 13 Kimi K2.6 fireworks 52.9%
  14. 14 Codestral openrouter 32.9%
  15. 15 Gemini 2.5 Pro Jun 2025, openrouter 5.80%

Source: Official Instrumentation Universal Knowledge v1.5 leaderboard

#EntryScoreSettingDateSource
1 Grok-3 mini 86.4% xai Apr 2025 Official leaderboard
2 Claude Opus 4.7 85.8% openrouter Apr 2026 Official leaderboard
3 Claude Sonnet 4.6 83.2% openrouter Feb 2026 Official leaderboard
4 Grok 4.3 83.2% openrouter Apr 2026 Official leaderboard
5 Devstral 2 71.4% openrouter Dec 2025 Official leaderboard
6 Gemma 3 27B 66% openrouter Mar 2025 Official leaderboard
7 Mistral Medium 3.5 65.9% openrouter Apr 2026 Official leaderboard
8 GPT-5.5 62.9% openrouter Apr 2026 Official leaderboard
9 Qwen 2.5 32B Instruct 60.1% ollama_remote Official leaderboard
10 Llama 4 Maverick 60% openrouter Apr 2025 Official leaderboard
11 Llama 4 Scout 57.6% openrouter Apr 2025 Official leaderboard
12 Gemma 3 12B 55.5% openrouter Mar 2025 Official leaderboard
13 Kimi K2.6 52.9% fireworks Apr 2026 Official leaderboard
14 Codestral 32.9% openrouter Jan 2025 Official leaderboard
15 Gemini 2.5 Pro Jun 2025 5.80% openrouter Jun 2025 Official leaderboard
#ModelScoreSettingDateSource
1 Grok-3 mini 86.4% xai Apr 2025 Official leaderboard
2 Claude Opus 4.7 85.8% openrouter Apr 2026 Official leaderboard
3 Claude Sonnet 4.6 83.2% openrouter Feb 2026 Official leaderboard
4 Grok 4.3 83.2% openrouter Apr 2026 Official leaderboard
5 Devstral 2 71.4% openrouter Dec 2025 Official leaderboard
6 Gemma 3 27B 66% openrouter Mar 2025 Official leaderboard
7 Mistral Medium 3.5 65.9% openrouter Apr 2026 Official leaderboard
8 GPT-5.5 62.9% openrouter Apr 2026 Official leaderboard
9 Llama 4 Maverick 60% openrouter Apr 2025 Official leaderboard
10 Llama 4 Scout 57.6% openrouter Apr 2025 Official leaderboard
11 Gemma 3 12B 55.5% openrouter Mar 2025 Official leaderboard
12 Kimi K2.6 52.9% fireworks Apr 2026 Official leaderboard
13 Codestral 32.9% openrouter Jan 2025 Official leaderboard
14 Gemini 2.5 Pro Jun 2025 5.80% openrouter Jun 2025 Official leaderboard

Rankings by part

We read 4 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as T1: Technician, T2: Engineer and T3: Specialist. Grok-3 mini leads 2 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

T1: Technician

14 models we track have a score here.

  1. 1 Claude Sonnet 4.6 ↑2 93.3%
  2. 2 Claude Opus 4.7 93.3%
  3. 3 GPT-5.5 ↑5 93.3%
  4. 4 Grok 4.3 93.3%
  5. 5 Grok-3 mini ↓4 90%
  6. 6 Llama 4 Maverick ↑3 85%
  7. 7 Kimi K2.6 ↑5 81.7%
  8. 8 Devstral 2 ↓3 80%
  9. 9 Gemma 3 27B ↓3 78.3%
  10. 10 Llama 4 Scout 76.7%

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

xAI
Own leaderboard Passed 85% in Jun 2026, by Grok-3 mini
Grok-3 mini 86.4%

Source: Official Instrumentation Universal Knowledge v1.5 leaderboard

Grok-3 mini on the official leaderboard passed 85% in April 2025

Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Grok-3 mini Gemma 3 27B Codestral 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Grok-3 mini Gemma 3 27B

Source: Official Instrumentation Universal Knowledge v1.5 leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

Versions

We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.

Anthropic xAI OpenAI
Instrumentation Universal Knowledge v1.6 refresh Passed 85% in Oct 2026, by GPT-6 Luna
GPT-6 Luna 93.3%
Instrumentation Universal Knowledge v1.3 Passed 85% in Jun 2026, by Claude Opus 4.7
Instrumentation Universal Knowledge v1.5 Passed 85% in Jun 2026, by Grok-3 mini
Grok-3 mini 86.4%
Instrumentation Universal Knowledge v1.6 refresh Passed 85% in Oct 2026, by GPT-6 Luna
GPT-6 Luna 93.3%

Source: Official leaderboards

More in Economic work

All 210 in Economic work
  • JobBench In use 130 tasks from 35 occupations workers most want delegated. Agent works through messy reference files, graded by fact-anchored binary rubrics. GLM-5.3 61.4%
  • GDPval In use Real work tasks from 44 occupations with reference files. Experts blindly compare model deliverables with expert work. Metric is win-or-tie rate. Seed2.1 Pro 87.9%
  • GDPval-AA v2.1 In use Agents with shell and web access complete 220 GDPval gold tasks from 44 occupations. Blind pairwise comparisons give Elo ratings. Claude Opus 5.5 1,846
  • AA-Briefcase v1.1 In use Version 1.1 of Artificial Analysis's multi-hour office work evaluation, with Elo ratings fitted by Crowd-BT. Claude Opus 5.5 1,822
  • Agents' Last Exam v1.1 In use Agents complete 122 public long-horizon professional workflow tasks across 55 sub-industries. Deterministic graders check outcomes against hidden references. GPT-6.2 Astra 47.5%
  • CoWorkBench In use Agent completes long-horizon productivity tasks across computer science, finance, law and medicine. Qwen in-house scoring (method unpublished). Claude Fable 5 75.9%

Data audit

All 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (15 entries).
  • Pass Models scored 14 models we track, plus 1 other entry on its leaderboard.
  • Note Newest score The newest came 7 months ago, in April 2026, for Grok 4.3. That’s expected for a test replaced by Instrumentation Universal Knowledge v1.6 refresh.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5.5 (#21).
  • Note Headroom Saturated. The best is 86.4%, past 85% since April 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on relayforge.tools: 14 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check