Benchmark Almanac

FlavourBench Complete Core

Grok 4.6 65.1 leads since August 2026
Not in model cards
Version complete core

Models choose three of eight ingredients for pairing, substitution, or a culinary constraint. A fixed scoring function assigns each possible combination a value from zero to 100.

The authors publish 27 model scores on the same 534 tasks. The scores measure agreement with Epicure’s ingredient rankings.

Released
August 2026
Kind
Taste and style
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Items from
Generated
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude Opus 5 #11 on the capability index
Last model added
August 2026
Scores refreshed
November 20th, 2026
27
Entries on its own leaderboard, 24 models we track

Grok 4.6 leads with 65.1 among models released by August 2026 Grok 4.6 leads with 65.1 among models released by August 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 August 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 4.6 minimal 65.1
  2. 2 Gemini 3.1 Pro minimal 65.0
  3. 3 GPT-5.6 Sol Pro minimal 64.2
  4. 4 Muse Spark 1.2 minimal 63.8
  5. 5 GPT-5.6 Terra Pro minimal 63.7
  6. 6 Claude Fable 5 minimal 63.4
  7. 7 GPT-5.6 Luna pro, minimal 62.6
  8. 8 Claude Opus 5 minimal 62.5
  9. 9 Qwen3.8-2.4T-A95B low 62.1
  10. 10 Kimi K3 low 62.0
  11. 11 Gemini 3.6 Flash minimal 62.0
  12. 12 DeepSeek V4 Pro 0813 minimal 62.0
  13. 13 Qwen 3.8 Max minimal 61.5
  14. 14 Tencent Hy3 minimal 61.5
  15. 15 MiniMax-M3 provider_fixed 60.9
  16. 16 GLM-5.3 provider_fixed 60.6
  17. 17 Muse Glimmer minimal 59.9
  18. 18 Seed 2.1 Turbo none 59.7
  19. 19 Inkling minimal 59.6
  20. 20 Claude Sonnet 5 minimal 59.5
  21. 21 GLM-5.2 none 58.5
  22. 22 Nemotron 3.5 Lightning 30B-A3B provider_fixed 57.4
  23. 23 Cohere Command A provider_fixed 56.7
  24. 24 DeepSeek V4 Flash 0731 none 55.4
  25. 25 Mistral Large 3 provider_fixed 55.4
  26. 26 Llama 4 Maverick provider_fixed 53.6
  27. 27 Command R+ (08-2024) provider_fixed 47.9

Source: Official FlavourBench Complete Core leaderboard

#EntryScore ±SettingDateSource
1 Grok 4.6 65.1 ±4.1 minimal, openrouter Aug 2026 Official leaderboard
2 Gemini 3.1 Pro 65.0 ±4.1 minimal, openrouter Feb 2026 Official leaderboard
3 GPT-5.6 Sol Pro 64.2 ±4.1 minimal, openrouter Jul 2026 Official leaderboard
4 Muse Spark 1.2 63.8 ±4.1 minimal, openrouter Aug 2026 Official leaderboard
5 GPT-5.6 Terra Pro 63.7 ±4.1 minimal, openrouter Official leaderboard
6 Claude Fable 5 63.4 ±4.2 minimal, openrouter Jun 2026 Official leaderboard
7 GPT-5.6 Luna 62.6 ±4.1 pro, minimal, openrouter Jul 2026 Official leaderboard
8 Claude Opus 5 62.5 ±4.1 minimal, openrouter Jul 2026 Official leaderboard
9 Qwen3.8-2.4T-A95B 62.1 ±4.2 low, openrouter Aug 2026 Official leaderboard
10 Kimi K3 62.0 ±4.1 low, openrouter Jul 2026 Official leaderboard
11 Gemini 3.6 Flash 62.0 ±4.3 minimal, openrouter Jul 2026 Official leaderboard
12 DeepSeek V4 Pro 0813 62.0 ±4.2 minimal, openrouter Aug 2026 Official leaderboard
13 Qwen 3.8 Max 61.5 ±4.1 minimal, openrouter Aug 2026 Official leaderboard
14 Tencent Hy3 61.5 ±4.2 minimal, openrouter Jul 2026 Official leaderboard
15 MiniMax-M3 60.9 ±4.2 provider_fixed, openrouter Jun 2026 Official leaderboard
16 GLM-5.3 60.6 ±4.2 provider_fixed, zai_coding_direct Aug 2026 Official leaderboard
17 Muse Glimmer 59.9 ±4 minimal, openrouter Aug 2026 Official leaderboard
18 Seed 2.1 Turbo 59.7 ±4.3 none, openrouter Official leaderboard
19 Inkling 59.6 ±4.2 minimal, openrouter Jul 2026 Official leaderboard
20 Claude Sonnet 5 59.5 ±4.1 minimal, openrouter Jun 2026 Official leaderboard
21 GLM-5.2 58.5 ±4.3 none, openrouter Jun 2026 Official leaderboard
22 Nemotron 3.5 Lightning 30B-A3B 57.4 ±4.2 provider_fixed, openrouter Aug 2026 Official leaderboard
23 Cohere Command A 56.7 ±4.2 provider_fixed, openrouter Mar 2025 Official leaderboard
24 DeepSeek V4 Flash 0731 55.4 ±4.3 none, openrouter Jul 2026 Official leaderboard
25 Mistral Large 3 55.4 ±4.1 provider_fixed, openrouter Dec 2025 Official leaderboard
26 Llama 4 Maverick 53.6 ±4 provider_fixed, openrouter Apr 2025 Official leaderboard
27 Command R+ (08-2024) 47.9 ±4.1 provider_fixed, openrouter Official leaderboard
#ModelScore ±SettingDateSource
1 Grok 4.6 65.1 ±4.1 minimal, openrouter Aug 2026 Official leaderboard
2 Gemini 3.1 Pro 65.0 ±4.1 minimal, openrouter Feb 2026 Official leaderboard
3 GPT-5.6 Sol Pro 64.2 ±4.1 minimal, openrouter Jul 2026 Official leaderboard
4 Muse Spark 1.2 63.8 ±4.1 minimal, openrouter Aug 2026 Official leaderboard
5 Claude Fable 5 63.4 ±4.2 minimal, openrouter Jun 2026 Official leaderboard
6 GPT-5.6 Luna 62.6 ±4.1 pro, minimal, openrouter Jul 2026 Official leaderboard
7 Claude Opus 5 62.5 ±4.1 minimal, openrouter Jul 2026 Official leaderboard
8 Qwen3.8-2.4T-A95B 62.1 ±4.2 low, openrouter Aug 2026 Official leaderboard
9 Kimi K3 62.0 ±4.1 low, openrouter Jul 2026 Official leaderboard
10 Gemini 3.6 Flash 62.0 ±4.3 minimal, openrouter Jul 2026 Official leaderboard
11 DeepSeek V4 Pro 0813 62.0 ±4.2 minimal, openrouter Aug 2026 Official leaderboard
12 Qwen 3.8 Max 61.5 ±4.1 minimal, openrouter Aug 2026 Official leaderboard
13 Tencent Hy3 61.5 ±4.2 minimal, openrouter Jul 2026 Official leaderboard
14 MiniMax-M3 60.9 ±4.2 provider_fixed, openrouter Jun 2026 Official leaderboard
15 GLM-5.3 60.6 ±4.2 provider_fixed, zai_coding_direct Aug 2026 Official leaderboard
16 Muse Glimmer 59.9 ±4 minimal, openrouter Aug 2026 Official leaderboard
17 Inkling 59.6 ±4.2 minimal, openrouter Jul 2026 Official leaderboard
18 Claude Sonnet 5 59.5 ±4.1 minimal, openrouter Jun 2026 Official leaderboard
19 GLM-5.2 58.5 ±4.3 none, openrouter Jun 2026 Official leaderboard
20 Nemotron 3.5 Lightning 30B-A3B 57.4 ±4.2 provider_fixed, openrouter Aug 2026 Official leaderboard
21 Cohere Command A 56.7 ±4.2 provider_fixed, openrouter Mar 2025 Official leaderboard
22 DeepSeek V4 Flash 0731 55.4 ±4.3 none, openrouter Jul 2026 Official leaderboard
23 Mistral Large 3 55.4 ±4.1 provider_fixed, openrouter Dec 2025 Official leaderboard
24 Llama 4 Maverick 53.6 ±4 provider_fixed, openrouter Apr 2025 Official leaderboard

Rankings by part

We read 5 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Constraint, Pairing and Substitution. Grok 4.6 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.

Constraint

24 models we track have a score here.

  1. 1 Grok 4.6 58.2
  2. 2 Muse Spark 1.2 ↑2 55.0
  3. 3 Gemini 3.1 Pro ↓1 54.2
  4. 4 GPT-5.6 Sol Pro ↓1 54.1
  5. 5 GPT-5.6 Luna ↑1 53.7
  6. 6 Qwen3.8-2.4T-A95B ↑2 52.5
  7. 7 Gemini 3.6 Flash ↑3 51.6
  8. 8 Claude Fable 5 ↓3 51.3
  9. 9 Tencent Hy3 ↑4 50.5
  10. 10 Claude Opus 5 ↓3 50.4

2 models have held number one on the official leaderboard since its release in August 2026

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

Google DeepMind xAI
Own leaderboard Replaced by Powered v43, Aug 2026
Grok 4.6 65.1

Source: Official FlavourBench Complete Core 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
55.0 60.0 65.0 2025 2026 55.0 60.0 65.0 2025 2026

Source: Official FlavourBench Complete Core 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

More in Writing and taste

All 65 in Writing and taste
  • Design Bench (Moonshot) In use Models make graphic designs from direct prompts or few-shot reference images. An automated grader rates the results as Elo scores. Kimi K3.1 Pro 957
  • WritingBench Fading Models complete 1,000 real-world writing requests across 6 domains and 100 subdomains. A judge model scores each reply on five query-specific criteria. GPT-5 83.9
  • Creative Writing (MiniMax in-house) Only in system cards Models write creative pieces for real Hailuo AI user requests. Automatic assessments score the writing. MiniMax-Text-01 81.3%
  • DeepSeek Chinese functional writing evaluation Only in system cards Models answer 3,170 everyday Chinese writing prompts such as reports, emails and speeches. Pairwise judging against a baseline model gives a win rate. DeepSeek-V4-Pro 62.6%
  • Slide generation human evaluation (Z.ai) Only in system cards People compare HTML slide decks from two models on content, layout and visual quality. Metric is overall win rate. GLM-5 67.5%
  • Task-dependent output homogenization evaluation Only in system cards Models answer prompts several times. A judge scores functional diversity where it is wanted and consistency where one answer is right. Muse Spark 0.21

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 (27 entries).
  • Pass Models scored 24 models we track, plus 3 other entries on its leaderboard.
  • Pass Newest score The newest is from August 2026, for GLM-5.3.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude Opus 5 (#11).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 24 rows for models we track and 3 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