Benchmark Almanac

Decodability Assessment Benchmark

GPT-5.5 95.9% leads since September 2026
Not in model cards

Models flag words a child cannot yet sound out in a passage, given the phonics lessons the child has completed. The leaderboard reports adjusted balanced accuracy.

Fab AI scores 32 model configurations on Kiswahili and 30 on English. Its public sample pairs one Kiswahili passage with two learner profiles.

Made by
Fab AI
Released
September 2026
Kind
Language reasoning
Level
University
Scored
Checked automatically
Answers
Short answer
Tasks
Public, plus a private test set
Measures
Capability
Language
Swahili, English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
A host
Strongest model tested
Claude Opus 5 #11 on the capability index
Last model added
July 2026
Scores refreshed
November 20th, 2026
32
Entries on its own leaderboard, 30 models we track

GPT-5.5 leads with 95.9% among models released by July 2026 GPT-5.5 leads with 95.9% among models released by July 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 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-5.5 95.9%
  2. 2 Claude Fable 5 95.9%
  3. 3 Gemini 3.6 Flash 95.7%
  4. 4 GPT-5.4 medium 95%
  5. 5 Gemini 3.5 Flash 94%
  6. 6 Claude Opus 5 92.7%
  7. 7 GPT-5.6 Sol 92.3%
  8. 8 Gemini 3 Flash 90.5%
  9. 9 GPT-5.6 Terra 89.8%
  10. 10 Grok 4.5 88.1%
  11. 11 Gemini 3.1 Pro 88%
  12. 12 GPT-5.6 Luna 84.1%
  13. 13 Claude Sonnet 5 78.3%
  14. 14 o3 66.2%
  15. 15 DeepSeek-V4-Pro 61.3%
  16. 16 gpt-oss-120b 55.8%
  17. 17 o4-mini 52.9%
  18. 18 GPT-5 nano medium 50.9%
  19. 19 Gemma 4 E2B 43.6%
  20. 20 Gemma 4 31B IT 33.2%
  21. 21 Gemma 4 12B 17.5%
  22. 22 Gemma 3 27B 15.4%
  23. 23 Ministral 3 8B 14.1%
  24. 24 Claude Haiku 4.5 12.1%
  25. 25 Ministral 3 3B 9.91%
  26. 26 Qwen3-1.7B 9.22%
  27. 27 Gemini 3.5 Flash-Lite 9.22%
  28. 28 GPT-4.1 mini 7.55%
  29. 29 Qwen3-4B 7.33%
  30. 30 GPT-4.1 nano 6.27%
  31. 31 Gemma 4 26B A4B 5.89%
  32. 32 Qwen3-0.6B 1.26%

Source: Official Decodability Assessment Benchmark leaderboard

#EntryScoreSettingDateSource
1 GPT-5.5 95.9% Apr 2026 Official leaderboard
2 Claude Fable 5 95.9% Jun 2026 Official leaderboard
3 Gemini 3.6 Flash 95.7% Jul 2026 Official leaderboard
4 GPT-5.4 95% medium Mar 2026 Official leaderboard
5 Gemini 3.5 Flash 94% May 2026 Official leaderboard
6 Claude Opus 5 92.7% Jul 2026 Official leaderboard
7 GPT-5.6 Sol 92.3% Jul 2026 Official leaderboard
8 Gemini 3 Flash 90.5% Dec 2025 Official leaderboard
9 GPT-5.6 Terra 89.8% Jul 2026 Official leaderboard
10 Grok 4.5 88.1% Jul 2026 Official leaderboard
11 Gemini 3.1 Pro 88% Feb 2026 Official leaderboard
12 GPT-5.6 Luna 84.1% Jul 2026 Official leaderboard
13 Claude Sonnet 5 78.3% Jun 2026 Official leaderboard
14 o3 66.2% Apr 2025 Official leaderboard
15 DeepSeek-V4-Pro 61.3% Apr 2026 Official leaderboard
16 gpt-oss-120b 55.8% Aug 2025 Official leaderboard
17 o4-mini 52.9% Apr 2025 Official leaderboard
18 GPT-5 nano 50.9% medium Aug 2025 Official leaderboard
19 Gemma 4 E2B 43.6% Mar 2026 Official leaderboard
20 Gemma 4 31B IT 33.2% Apr 2026 Official leaderboard
21 Gemma 4 12B 17.5% Jun 2026 Official leaderboard
22 Gemma 3 27B 15.4% Mar 2025 Official leaderboard
23 Ministral 3 8B 14.1% Official leaderboard
24 Claude Haiku 4.5 12.1% Oct 2025 Official leaderboard
25 Ministral 3 3B 9.91% Official leaderboard
26 Qwen3-1.7B 9.22% Apr 2025 Official leaderboard
27 Gemini 3.5 Flash-Lite 9.22% Jul 2026 Official leaderboard
28 GPT-4.1 mini 7.55% Apr 2025 Official leaderboard
29 Qwen3-4B 7.33% Apr 2025 Official leaderboard
30 GPT-4.1 nano 6.27% Apr 2025 Official leaderboard
31 Gemma 4 26B A4B 5.89% Apr 2026 Official leaderboard
32 Qwen3-0.6B 1.26% Apr 2025 Official leaderboard
#ModelScoreSettingDateSource
1 GPT-5.5 95.9% Apr 2026 Official leaderboard
2 Claude Fable 5 95.9% Jun 2026 Official leaderboard
3 Gemini 3.6 Flash 95.7% Jul 2026 Official leaderboard
4 GPT-5.4 95% medium Mar 2026 Official leaderboard
5 Gemini 3.5 Flash 94% May 2026 Official leaderboard
6 Claude Opus 5 92.7% Jul 2026 Official leaderboard
7 GPT-5.6 Sol 92.3% Jul 2026 Official leaderboard
8 Gemini 3 Flash 90.5% Dec 2025 Official leaderboard
9 GPT-5.6 Terra 89.8% Jul 2026 Official leaderboard
10 Grok 4.5 88.1% Jul 2026 Official leaderboard
11 Gemini 3.1 Pro 88% Feb 2026 Official leaderboard
12 GPT-5.6 Luna 84.1% Jul 2026 Official leaderboard
13 Claude Sonnet 5 78.3% Jun 2026 Official leaderboard
14 o3 66.2% Apr 2025 Official leaderboard
15 DeepSeek-V4-Pro 61.3% Apr 2026 Official leaderboard
16 gpt-oss-120b 55.8% Aug 2025 Official leaderboard
17 o4-mini 52.9% Apr 2025 Official leaderboard
18 GPT-5 nano 50.9% medium Aug 2025 Official leaderboard
19 Gemma 4 E2B 43.6% Mar 2026 Official leaderboard
20 Gemma 4 31B IT 33.2% Apr 2026 Official leaderboard
21 Gemma 4 12B 17.5% Jun 2026 Official leaderboard
22 Gemma 3 27B 15.4% Mar 2025 Official leaderboard
23 Claude Haiku 4.5 12.1% Oct 2025 Official leaderboard
24 Qwen3-1.7B 9.22% Apr 2025 Official leaderboard
25 Gemini 3.5 Flash-Lite 9.22% Jul 2026 Official leaderboard
26 GPT-4.1 mini 7.55% Apr 2025 Official leaderboard
27 Qwen3-4B 7.33% Apr 2025 Official leaderboard
28 GPT-4.1 nano 6.27% Apr 2025 Official leaderboard
29 Gemma 4 26B A4B 5.89% Apr 2026 Official leaderboard
30 Qwen3-0.6B 1.26% Apr 2025 Official leaderboard

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 Sep 2026, by GPT-5.5
GPT-5.5 95.9%

Source: Official Decodability Assessment Benchmark leaderboard

Gemini 3 Flash on the official leaderboard passed 85% in December 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%) GPT-5.5 Gemini 3 Flash o3 Gemma 3 27B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) GPT-5.5 o3

Source: Official Decodability Assessment 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 Knowledge

All 332 in Knowledge
  • MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects. Intern-S2-397B 89.8%
  • MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy. Seed1.8 92.3%
  • SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy. Qwen 3.6 Max (Preview) 73.9%
  • WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1. Tencent Hy4 preview 83.9%
  • BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly. Gemma 4 31B IT 74.4%
  • HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (32 entries).
  • Pass Models scored 30 models we track, plus 2 other entries on its leaderboard.
  • Pass Newest score The newest is from July 2026, for Claude Opus 5.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Claude Opus 5 (#11).
  • Note Headroom Saturated. The best is 95.9%, past 85% since December 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on fab-ai.org: 30 rows for models we track and 2 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