Benchmark Almanac

Instrumentation Universal Knowledge v1.6 refresh

GPT-6 Luna 93.3% leads since October 2026
Not in model cards
Version 1.6

Models answer 60 technician questions and 97 specialist questions from the June samples. Three model judges score responses under higher output limits.

RelayForge ranks 31 models using a two-tier composite. Eight results are provisional. The instructor has yet to review the automated judgments.

Made by
RelayForge
Released
October 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
Claude Opus 5.5 #5 on the capability index
Last model added
October 2026
Scores refreshed
November 20th, 2026
31
Entries on its own leaderboard, 25 models we track

GPT-6 Luna leads with 93.3% GPT-6 Luna leads with 93.3%

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.

  1. 1 GPT-6 Luna 93.3%
  2. 2 Claude Sonnet 5.5 91.8%
  3. 3 Claude Opus 4.7 91.7%
  4. 4 GPT-5.5 87.3%
  5. 5 Claude Sonnet 4.6 85.7%
  6. 6 Gemini 4 Argon direct google API 85.7%
  7. 7 Gemini 2.5 Pro Jun 2025 83.7%
  8. 8 Grok 4.3 83.7%
  9. 9 Gemini 3.8 Flash direct google API 83.3%
  10. 10 Claude Opus 5.5 direct anthropic API 83.3%
  11. 11 GLM-5.3 82.2%
  12. 12 Claude Haiku 5.5 direct anthropic API 81.7%
  13. 13 Grok 4.7 81.4%
  14. 14 Devstral 2 (2512) 77.9%
  15. 15 Gemma 4 31B (local QAT) local 76.6%
  16. 16 Qwen 3.8 27B re-run pending 75.8%
  17. 17 Llama 4 Maverick 74.6%
  18. 18 Gemma 3 27B 69.1%
  19. 19 gpt-oss-120b 68.7%
  20. 20 Gemma 4 26B-A4B (local QAT) provisional, local 64.6%
  21. 21 DeepSeek V4.1 Flash provisional, re-run pending 63.4%
  22. 22 DeepSeek-V4-Pro provisional, re-run pending 62.1%
  23. 23 Qwen 2.5 32B (local) local 61.5%
  24. 24 Kimi K2.6 provisional 61.3%
  25. 25 Gemma 3 12B 54.1%
  26. 26 Llama 4 Scout 53.3%
  27. 27 Mistral Large 4 provisional, re-run pending 51.3%
  28. 28 Codestral (2508) 48.2%
  29. 29 Mistral Medium 3.5 provisional, re-run pending 43.2%
  30. 30 Qwen 3.6 35B-A3B provisional, re-run pending 40.6%
  31. 31 Mistral Small (2603) provisional, re-run pending 29.2%

Source: Official Instrumentation Universal Knowledge v1.6 refresh leaderboard

#EntryScoreSettingDateSource
1 GPT-6 Luna 93.3% Sep 2026 Official leaderboard
2 Claude Sonnet 5.5 91.8% Sep 2026 Official leaderboard
3 Claude Opus 4.7 91.7% Apr 2026 Official leaderboard
4 GPT-5.5 87.3% Apr 2026 Official leaderboard
5 Claude Sonnet 4.6 85.7% Feb 2026 Official leaderboard
6 Gemini 4 Argon 85.7% direct google API Sep 2026 Official leaderboard
7 Gemini 2.5 Pro Jun 2025 83.7% Jun 2025 Official leaderboard
8 Grok 4.3 83.7% Apr 2026 Official leaderboard
9 Gemini 3.8 Flash 83.3% direct google API Sep 2026 Official leaderboard
10 Claude Opus 5.5 83.3% direct anthropic API Sep 2026 Official leaderboard
11 GLM-5.3 82.2% Aug 2026 Official leaderboard
12 Claude Haiku 5.5 81.7% direct anthropic API Oct 2026 Official leaderboard
13 Grok 4.7 81.4% Sep 2026 Official leaderboard
14 Devstral 2 (2512) 77.9% Official leaderboard
15 Gemma 4 31B (local QAT) 76.6% local Official leaderboard
16 Qwen 3.8 27B 75.8% re-run pending Aug 2026 Official leaderboard
17 Llama 4 Maverick 74.6% Apr 2025 Official leaderboard
18 Gemma 3 27B 69.1% Mar 2025 Official leaderboard
19 gpt-oss-120b 68.7% Aug 2025 Official leaderboard
20 Gemma 4 26B-A4B (local QAT) 64.6% provisional, local Official leaderboard
21 DeepSeek V4.1 Flash 63.4% provisional, re-run pending Sep 2026 Official leaderboard
22 DeepSeek-V4-Pro 62.1% provisional, re-run pending Apr 2026 Official leaderboard
23 Qwen 2.5 32B (local) 61.5% local Official leaderboard
24 Kimi K2.6 61.3% provisional Apr 2026 Official leaderboard
25 Gemma 3 12B 54.1% Mar 2025 Official leaderboard
26 Llama 4 Scout 53.3% Apr 2025 Official leaderboard
27 Mistral Large 4 51.3% provisional, re-run pending Oct 2026 Official leaderboard
28 Codestral (2508) 48.2% Official leaderboard
29 Mistral Medium 3.5 43.2% provisional, re-run pending Apr 2026 Official leaderboard
30 Qwen 3.6 35B-A3B 40.6% provisional, re-run pending Apr 2026 Official leaderboard
31 Mistral Small (2603) 29.2% provisional, re-run pending Official leaderboard
#ModelScoreSettingDateSource
1 GPT-6 Luna 93.3% Sep 2026 Official leaderboard
2 Claude Sonnet 5.5 91.8% Sep 2026 Official leaderboard
3 Claude Opus 4.7 91.7% Apr 2026 Official leaderboard
4 GPT-5.5 87.3% Apr 2026 Official leaderboard
5 Claude Sonnet 4.6 85.7% Feb 2026 Official leaderboard
6 Gemini 4 Argon 85.7% direct google API Sep 2026 Official leaderboard
7 Gemini 2.5 Pro Jun 2025 83.7% Jun 2025 Official leaderboard
8 Grok 4.3 83.7% Apr 2026 Official leaderboard
9 Gemini 3.8 Flash 83.3% direct google API Sep 2026 Official leaderboard
10 Claude Opus 5.5 83.3% direct anthropic API Sep 2026 Official leaderboard
11 GLM-5.3 82.2% Aug 2026 Official leaderboard
12 Claude Haiku 5.5 81.7% direct anthropic API Oct 2026 Official leaderboard
13 Grok 4.7 81.4% Sep 2026 Official leaderboard
14 Qwen 3.8 27B 75.8% re-run pending Aug 2026 Official leaderboard
15 Llama 4 Maverick 74.6% Apr 2025 Official leaderboard
16 Gemma 3 27B 69.1% Mar 2025 Official leaderboard
17 gpt-oss-120b 68.7% Aug 2025 Official leaderboard
18 DeepSeek V4.1 Flash 63.4% provisional, re-run pending Sep 2026 Official leaderboard
19 DeepSeek-V4-Pro 62.1% provisional, re-run pending Apr 2026 Official leaderboard
20 Kimi K2.6 61.3% provisional Apr 2026 Official leaderboard
21 Gemma 3 12B 54.1% Mar 2025 Official leaderboard
22 Llama 4 Scout 53.3% Apr 2025 Official leaderboard
23 Mistral Large 4 51.3% provisional, re-run pending Oct 2026 Official leaderboard
24 Mistral Medium 3.5 43.2% provisional, re-run pending Apr 2026 Official leaderboard
25 Qwen 3.6 35B-A3B 40.6% provisional, re-run pending Apr 2026 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: T1 and T3. Each ranks models on one part of the test. Gemini 4 Argon leads T1 and GPT-6 Luna leads T3. Arrows show how far a model moved from its overall rank among the models we track.

T1

25 models we track have a score here.

  1. 1 Gemini 4 Argon ↑5 96.7%
  2. 2 Claude Opus 4.7 ↑1 95%
  3. 3 GPT-6 Luna ↓2 95%
  4. 4 GPT-5.5 93.3%
  5. 5 Gemini 3.8 Flash ↑4 93.3%
  6. 6 Claude Opus 5.5 ↑4 93.3%
  7. 7 Claude Haiku 5.5 ↑5 93.3%
  8. 8 Gemini 2.5 Pro Jun 2025 ↓1 91.7%
  9. 9 Kimi K2.6 ↑11 91.7%
  10. 10 Grok 4.3 ↓2 91.7%

Number one over time

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

OpenAI
Own leaderboard Passed 85% in Oct 2026, by GPT-6 Luna
GPT-6 Luna 93.3%

Source: Official Instrumentation Universal Knowledge v1.6 refresh leaderboard

Claude Sonnet 4.6 on the official leaderboard passed 85% in February 2026

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%) GPT-6 Luna Claude Opus 4.7 Claude Sonnet 4.6 Gemini 2.5 Pro Llama 4 Maverick Gemma 3 27B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) GPT-6 Luna Claude Sonnet 4.6

Source: Official Instrumentation Universal Knowledge v1.6 refresh 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 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (31 entries).
  • Pass Models scored 25 models we track, plus 6 other entries on its leaderboard.
  • Pass Newest score The newest is from October 2026, for Claude Haiku 5.5.
  • Pass Frontier models 3 of the top 10 have a score, led by Claude Opus 5.5 (#5).
  • Note Headroom Saturated. The best is 93.3%, past 85% since February 2026.
Leaderboard import 3 pass
  • Pass Importer We read the board on relayforge.tools: 25 rows for models we track and 6 other entries 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