Benchmark Almanac

LLM Position Bias Benchmark

MiMo-V2-Pro 19.8% leads since March 2026
Not in model cards

Models judge pairs of responses twice, with the presentation order reversed.

The author publishes the model comparison in the repository.

Made by
Lech Mazur
Kind
Bias, fairness and politics
Level
Everyday
Scored
Checked automatically
Answers
Prediction
Tasks
All public
Measures
Propensity
Behavior
Unreliability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude Fable 5 #12 on the capability index
Last model added
June 2026
Scores refreshed
November 20th, 2026
31
Models scored on its own leaderboard

MiMo-V2-Pro leads with 19.8% among models released by June 2026

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.

The newest model with a score here came out in June 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 MiMo-V2-Pro 19.8%
  2. 2 Claude Fable 5 medium 27.3%
  3. 3 Claude Opus 4.8 high 27.4%
  4. 4 Seed2.0 Pro 28%
  5. 5 Gemini 3.5 Flash 29.8%
  6. 6 Claude Opus 4.6 high 30.2%
  7. 7 DeepSeek-V3.2 30.3%
  8. 8 GLM-5.1 31.5%
  9. 9 Qwen 3.6 Plus 33.8%
  10. 10 Qwen3.7-Max 34.8%
  11. 11 MiniMax-M3 34.9%
  12. 12 Gemini 3.1 Pro 35.4%
  13. 13 MiniMax-M2.7 36.5%
  14. 14 Trinity Large Thinking 36.6%
  15. 15 Claude Sonnet 4.6 high 37.4%
  16. 16 Claude Opus 4.7 high 37.7%
  17. 17 Grok 4.20 reasoning 39%
  18. 18 DeepSeek-V4-Pro 43.6%
  19. 19 GPT-5.5 medium 44.4%
  20. 20 ERNIE 5.0 45%
  21. 21 Gemini 3.1 Flash-Lite 45.3%
  22. 22 Kimi K2.6 47.3%
  23. 23 Qwen3.5 397B-A17B 48.5%
  24. 24 Gemma 4 31B IT 50.8%
  25. 25 Kimi K2.5 thinking 51.6%
  26. 26 Qwen3.5-122B-A10B 51.7%
  27. 27 GPT-5.4 Mini xhigh 54.7%
  28. 28 GPT-5.4 no reasoning 57.4%
  29. 29 Mistral Medium 3.1 57.9%
  30. 30 Mistral Large 3 58.7%
  31. 31 Mistral Medium 3.5 72.5%

Source: Official LLM Position Bias Benchmark leaderboard

#ModelScoreSettingDateSource
1 MiMo-V2-Pro 19.8% Mar 2026 Official leaderboard
2 Claude Fable 5 27.3% medium Jun 2026 Official leaderboard
3 Claude Opus 4.8 27.4% high May 2026 Official leaderboard
4 Seed2.0 Pro 28% Feb 2026 Official leaderboard
5 Gemini 3.5 Flash 29.8% May 2026 Official leaderboard
6 Claude Opus 4.6 30.2% high Feb 2026 Official leaderboard
7 DeepSeek-V3.2 30.3% Dec 2025 Official leaderboard
8 GLM-5.1 31.5% Apr 2026 Official leaderboard
9 Qwen 3.6 Plus 33.8% Mar 2026 Official leaderboard
10 Qwen3.7-Max 34.8% May 2026 Official leaderboard
11 MiniMax-M3 34.9% Jun 2026 Official leaderboard
12 Gemini 3.1 Pro 35.4% Feb 2026 Official leaderboard
13 MiniMax-M2.7 36.5% Mar 2026 Official leaderboard
14 Trinity Large Thinking 36.6% Apr 2026 Official leaderboard
15 Claude Sonnet 4.6 37.4% high Feb 2026 Official leaderboard
16 Claude Opus 4.7 37.7% high Apr 2026 Official leaderboard
17 Grok 4.20 39% reasoning Feb 2026 Official leaderboard
18 DeepSeek-V4-Pro 43.6% Apr 2026 Official leaderboard
19 GPT-5.5 44.4% medium Apr 2026 Official leaderboard
20 ERNIE 5.0 45% Jan 2026 Official leaderboard
21 Gemini 3.1 Flash-Lite 45.3% Mar 2026 Official leaderboard
22 Kimi K2.6 47.3% Apr 2026 Official leaderboard
23 Qwen3.5 397B-A17B 48.5% Feb 2026 Official leaderboard
24 Gemma 4 31B IT 50.8% Apr 2026 Official leaderboard
25 Kimi K2.5 51.6% thinking Jan 2026 Official leaderboard
26 Qwen3.5-122B-A10B 51.7% Feb 2026 Official leaderboard
27 GPT-5.4 Mini 54.7% xhigh Mar 2026 Official leaderboard
28 GPT-5.4 57.4% no reasoning Mar 2026 Official leaderboard
29 Mistral Medium 3.1 57.9% Aug 2025 Official leaderboard
30 Mistral Large 3 58.7% Dec 2025 Official leaderboard
31 Mistral Medium 3.5 72.5% Apr 2026 Official leaderboard

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
GPT-5.4
no reasoning high
66.3%
Gemma 4 31B IT
reasoning
54.9%
Grok 4.20
reasoning non-reasoni…
64.8%
Claude Sonnet 4.6
no reasoning high
44.8%
Claude Opus 4.6
no reasoning high
38.1%

Source: Official LLM Position Bias Benchmark leaderboard

5 models have held number one on the official leaderboard since 2023

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

Mistral AI DeepSeek Anthropic ByteDance Seed Xiaomi Corp
Own leaderboard Not tested since Jun 2026
MiMo-V2-Pro 19.8%

Source: Official LLM Position Bias Benchmark leaderboard

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
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 MiMo-V2-Pro Seed2.0 Pro DeepSeek-V3.2 Mistral Medium 3.1 0% 20% 40% 60% 80% 100% 2026 MiMo-V2-Pro Seed2.0 Pro

Source: Official LLM Position Bias Benchmark 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 Safety and alignment

All 636 in Safety and alignment
  • FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
  • StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is. Inkling 98.6%
  • Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. Gemini 4 Carbon 0%
  • Covert misbehavior (Thinking Machines) In use Agentic tasks that check whether a model misbehaves and conceals it. Scored by the rate of covert misbehavior. Inkling 2 Mini 8.30%
  • Destructive actions (Thinking Machines) In use Agentic tasks that check whether a model takes destructive actions it was not authorized to take. Scored by the rate of destructive actions. Inkling 2 Small 4.70%
  • Misbehavior (Thinking Machines) In use Agentic tasks that check whether a model fabricates results, ignores instructions or hides what it did when a task is hard. Scored by the rate of misbehavior. Inkling 2 Mini 9.80%

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (31 entries).
  • Pass Models scored 31 models we track.
  • Pass Newest score The newest is from June 2026, for Claude Fable 5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude Fable 5 (#12).
  • Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 36 rows for models we track 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