Benchmark Almanac

ManyIH-Bench

GPT-5.4 60.9% leads since April 2026
Not in model cards

Models resolve conflicts among instructions at up to twelve privilege levels. A response must satisfy all winning instructions and any applicable coding tests.

The authors publish 853 tasks. Prime Intellect independently ran the instruction-following subset in six model and harness combinations.

Released
April 2026
Kind
Following precise instructions
Level
University
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Text, Code
Interaction
Single turn
Environment
Code sandbox
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude Opus 5 #11 on the capability index
Last model added
July 2026
Scores refreshed
November 21st, 2026
13
Models scored in its paper

GPT-5.4 leads with 60.9% among models released by July 2026

Best score per model from the benchmark's paper, read on November 21st, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best.

The newest model with a score here came out in July 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 GPT-5.4 60.9%
  2. 2 Gemini 3.1 Pro 59%
  3. 3 Grok 4.20 54.1%
  4. 4 GPT-5.5 53.9%
  5. 5 GPT-5.6 Sol 52.5%
  6. 6 Claude Opus 4.6 51.3%
  7. 7 Claude Opus 5 50.4%
  8. 8 Kimi K2.5 42.4%
  9. 9 Qwen3.5 397B-A17B 41%
  10. 10 Claude Sonnet 4.6 39.1%
  11. 11 Qwen3.5-122B-A10B 19.7%
  12. 12 Qwen3.5-9B 8.40%
  13. 13 Qwen3.5-4B 3.50%

Source: The ManyIH-Bench paper

#ModelScoreSettingDateSource
1 GPT-5.4 60.9% Mar 2026 Official leaderboard
2 Gemini 3.1 Pro 59% Feb 2026 Official leaderboard
3 Grok 4.20 54.1% Feb 2026 Official leaderboard
4 GPT-5.5 53.9% Apr 2026 Official leaderboard
5 GPT-5.6 Sol 52.5% Jul 2026 Official leaderboard
6 Claude Opus 4.6 51.3% Feb 2026 Official leaderboard
7 Claude Opus 5 50.4% Jul 2026 Official leaderboard
8 Kimi K2.5 42.4% Jan 2026 Official leaderboard
9 Qwen3.5 397B-A17B 41% Feb 2026 Official leaderboard
10 Claude Sonnet 4.6 39.1% Feb 2026 Official leaderboard
11 Qwen3.5-122B-A10B 19.7% Feb 2026 Official leaderboard
12 Qwen3.5-9B 8.40% Feb 2026 Official leaderboard
13 Qwen3.5-4B 3.50% Feb 2026 Official leaderboard

Number one over time

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

OpenAI
Benchmark paper Not tested since Jul 2026
GPT-5.4 60.9%

Source: The ManyIH-Bench paper

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) GPT-5.4 Grok 4.20 Kimi K2.5 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) GPT-5.4 Grok 4.20

Source: Official ManyIH-Bench 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.

More in Instructions and chat

All 144 in Instructions and chat
  • IFEval In use Models answer about 500 prompts containing verifiable instructions such as word limits or required keywords. Code checks each instruction for strict and loose accuracy. Gemma 4 31B IT 98.9%
  • IFBench In use 58 new out-of-domain verifiable output constraints. Scored by programmatic constraint checks (strict or loose). Flint 89.2%
  • MultiChallenge In use Multi-turn conversations testing instruction retention, inference memory, versioned editing and self-coherence. Graded by LLM judge with rubrics. GPT-6.1 Sol 82%
  • Multi-IF In use Models follow verifiable instructions over three-turn conversations in eight languages, extending IFEval. Scored by instruction accuracy at each turn. LFM3-1.7B 89.1%
  • AdvancedIF In use Models respond to over 1,600 prompts with complex, multi-turn, and system-level instructions. A judge model checks each response against expert-written rubrics. MAI-Thinking-1 85%
  • MM-IFEval In use Multimodal models follow image-grounded instructions with exact output constraints. Checks mix rule-based verification and a judge model. GPT-4o 64.6%

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 benchmark’s paper (13 entries).
  • Pass Models scored 13 models we track.
  • Pass Newest score The newest is from July 2026, for Claude Opus 5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude Opus 5 (#11).
  • Pass Headroom The best is 60.9%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 13 rows for models we track at the last read.
  • Pass Last read November 21st, 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