Benchmark Almanac

OGS-Bench

GPT-5.5 86.8 leads since April 2026
Not in model cards

Models use a supplied culinary vocabulary to choose food and drink pairings, classify dishes, and produce structured answers. The headline score combines automatically checked tasks.

Gastrona ranks 20 models on a shared set of 213 tasks. The headline score excludes tasks graded by an unvalidated model judge.

Made by
Gastrona
Released
April 2026
Kind
Language reasoning
Level
University
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
20
Entries on its own leaderboard, 17 models we track

GPT-5.5 leads with 86.8 among models released by April 2026 GPT-5.5 leads with 86.8 among models released by April 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. Every run is in the next chart. 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 April 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 Open book 86.8
  2. 2 Claude Opus 4.7 Open book 85.9
  3. 3 GPT-5.4 Open book 85.3
  4. 4 Gemini 2.5 Pro Jun 2025, Open book 85.0
  5. 5 GPT-4o Mar 2025, Open book 80.9
  6. 6 Gemini 2.5 Flash Sep 2025, Open book 80.8
  7. 7 GPT-4.1 Open book 80.5
  8. 8 Claude Sonnet 4.6 Open book 80.3
  9. 9 GPT-5.4 Mini Open book 79.6
  10. 10 grok-4.20-0309-non-reasoning Open book 79.2
  11. 11 MiniMax-M2.7-highspeed Open book 79.0
  12. 12 Claude Haiku 4.5 Open book 78.7
  13. 13 Grok 4 Fast Open book 76.4
  14. 14 GPT-4.1 mini Open book 76.4
  15. 15 Gemini 2.5 Flash-Lite Sep 2025, Open book 75.1
  16. 16 GPT-4o mini Open book 72.8
  17. 17 GPT-5.4 Nano Open book 71.3
  18. 18 mistral-small-latest Open book 71.1
  19. 19 GPT-3.5 Turbo Jan 2024, Open book 62.1
  20. 20 GPT-4.1 nano Open book 61.9

Source: Official OGS-Bench leaderboard

#EntryScore ±SettingReleasedRun onLinkSource
1 GPT-5.5 86.8 ±3.2 Open book Apr 2026 May 2026 opengastronomy.org Official leaderboard
2 Claude Opus 4.7 85.9 ±3.6 Open book Apr 2026 Apr 2026 opengastronomy.org Official leaderboard
3 GPT-5.4 85.3 ±3.6 Open book Mar 2026 Apr 2026 opengastronomy.org Official leaderboard
4 Gemini 2.5 Pro Jun 2025 85.0 ±3 Open book Jun 2025 Apr 2026 opengastronomy.org Official leaderboard
5 GPT-4o Mar 2025 80.9 ±3.6 Open book Mar 2025 Apr 2026 opengastronomy.org Official leaderboard
6 Gemini 2.5 Flash Sep 2025 80.8 ±3.8 Open book Sep 2025 Apr 2026 opengastronomy.org Official leaderboard
7 GPT-4.1 80.5 ±4 Open book Apr 2025 Apr 2026 opengastronomy.org Official leaderboard
8 Claude Sonnet 4.6 80.3 ±4 Open book Feb 2026 Apr 2026 opengastronomy.org Official leaderboard
9 GPT-5.4 Mini 79.6 ±3.9 Open book Mar 2026 Apr 2026 opengastronomy.org Official leaderboard
10 grok-4.20-0309-non-reasoning 79.2 ±3.8 Open book Apr 2026 opengastronomy.org Official leaderboard
11 MiniMax-M2.7-highspeed 79.0 ±4.6 Open book May 2026 opengastronomy.org Official leaderboard
12 Claude Haiku 4.5 78.7 ±4.1 Open book Oct 2025 Apr 2026 opengastronomy.org Official leaderboard
13 Grok 4 Fast 76.4 ±4.6 Open book Sep 2025 Apr 2026 opengastronomy.org Official leaderboard
14 GPT-4.1 mini 76.4 ±4.1 Open book Apr 2025 Apr 2026 opengastronomy.org Official leaderboard
15 Gemini 2.5 Flash-Lite Sep 2025 75.1 ±4.4 Open book Sep 2025 Apr 2026 opengastronomy.org Official leaderboard
16 GPT-4o mini 72.8 ±4.3 Open book Jul 2024 Apr 2026 opengastronomy.org Official leaderboard
17 GPT-5.4 Nano 71.3 ±4.4 Open book Mar 2026 Apr 2026 opengastronomy.org Official leaderboard
18 mistral-small-latest 71.1 ±4.5 Open book Apr 2026 opengastronomy.org Official leaderboard
19 GPT-3.5 Turbo Jan 2024 62.1 ±5 Open book Jan 2024 Apr 2026 opengastronomy.org Official leaderboard
20 GPT-4.1 nano 61.9 ±5.4 Open book Apr 2025 Apr 2026 opengastronomy.org Official leaderboard
#ModelScore ±SettingReleasedRun onLinkSource
1 GPT-5.5 86.8 ±3.2 Open book Apr 2026 May 2026 opengastronomy.org Official leaderboard
2 Claude Opus 4.7 85.9 ±3.6 Open book Apr 2026 Apr 2026 opengastronomy.org Official leaderboard
3 GPT-5.4 85.3 ±3.6 Open book Mar 2026 Apr 2026 opengastronomy.org Official leaderboard
4 Gemini 2.5 Pro Jun 2025 85.0 ±3 Open book Jun 2025 Apr 2026 opengastronomy.org Official leaderboard
5 GPT-4o Mar 2025 80.9 ±3.6 Open book Mar 2025 Apr 2026 opengastronomy.org Official leaderboard
6 Gemini 2.5 Flash Sep 2025 80.8 ±3.8 Open book Sep 2025 Apr 2026 opengastronomy.org Official leaderboard
7 GPT-4.1 80.5 ±4 Open book Apr 2025 Apr 2026 opengastronomy.org Official leaderboard
8 Claude Sonnet 4.6 80.3 ±4 Open book Feb 2026 Apr 2026 opengastronomy.org Official leaderboard
9 GPT-5.4 Mini 79.6 ±3.9 Open book Mar 2026 Apr 2026 opengastronomy.org Official leaderboard
10 Claude Haiku 4.5 78.7 ±4.1 Open book Oct 2025 Apr 2026 opengastronomy.org Official leaderboard
11 Grok 4 Fast 76.4 ±4.6 Open book Sep 2025 Apr 2026 opengastronomy.org Official leaderboard
12 GPT-4.1 mini 76.4 ±4.1 Open book Apr 2025 Apr 2026 opengastronomy.org Official leaderboard
13 Gemini 2.5 Flash-Lite Sep 2025 75.1 ±4.4 Open book Sep 2025 Apr 2026 opengastronomy.org Official leaderboard
14 GPT-4o mini 72.8 ±4.3 Open book Jul 2024 Apr 2026 opengastronomy.org Official leaderboard
15 GPT-5.4 Nano 71.3 ±4.4 Open book Mar 2026 Apr 2026 opengastronomy.org Official leaderboard
16 GPT-3.5 Turbo Jan 2024 62.1 ±5 Open book Jan 2024 Apr 2026 opengastronomy.org Official leaderboard
17 GPT-4.1 nano 61.9 ±5.4 Open book Apr 2025 Apr 2026 opengastronomy.org Official leaderboard

Of runs with a cost, the best scores 85.9 for $16 for the full run

Runs on the cost frontier score higher than any run that's cheaper. GPT-5.4 at Open book comes within 0.6 of the best for $1.01 for the full run, 16 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

xAI Google DeepMind OpenAI Anthropic Cost frontier One model at each effort level
0.00 20.0 40.0 60.0 80.0 100.0 $0.01 $0.10 $1.00 $10 $100 Cost of the full run, log scale Claude Opus 4.7, Open book GPT-5.4, Open book Gemini 2.5 Flash, Open book Grok 4 Fast, Open book GPT-5.4 Nano, Open book 0.00 20.0 40.0 60.0 80.0 100.0 $0.01 $0.10 $1.00 $10 $100 Cost of the full run, log scale Claude Opus 4.7, Open book GPT-5.4 Mini, Open book

Source: Official OGS-Bench leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-5.5 Open book 86.8 163
Claude Opus 4.7 Open book 85.9 $16 92
GPT-5.4 Open book 85.3 $1.01 67
Gemini 2.5 Pro Jun 2025 Open book 85.0 $1.03 62
GPT-4o Mar 2025 Open book 80.9 $1.47 48
Gemini 2.5 Flash Sep 2025 Open book 80.8 $0.25 65
GPT-4.1 Open book 80.5 $1.27 66
Claude Sonnet 4.6 Open book 80.3 $2.73 110
GPT-5.4 Mini Open book 79.6 $0.18 50
Claude Haiku 4.5 Open book 78.7 $0.66 82
Grok 4 Fast Open book 76.4 $0.13 59
GPT-4.1 mini Open book 76.4 $0.24 54
Gemini 2.5 Flash-Lite Sep 2025 Open book 75.1 $0.06 54
GPT-4o mini Open book 72.8 $0.08 35
GPT-5.4 Nano Open book 71.3 $0.04 80
GPT-3.5 Turbo Jan 2024 Open book 62.1 $0.26 28
GPT-4.1 nano Open book 61.9 $0.06 40

Of runs with a cost, the best scores 85.9 for $16 for the full run

Runs on the cost frontier score higher than any run that's cheaper. GPT-5.4 at Open book comes within 0.6 of the best for $1.01 for the full run, 16 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

xAI Google DeepMind OpenAI Anthropic Cost frontier One model at each effort level
0.00 20.0 40.0 60.0 80.0 100.0 $0.01 $0.10 $1.00 $10 $100 Cost of the full run, log scale Claude Opus 4.7, Open book GPT-5.4, Open book Gemini 2.5 Flash, Open book Grok 4 Fast, Open book GPT-5.4 Nano, Open book 0.00 20.0 40.0 60.0 80.0 100.0 $0.01 $0.10 $1.00 $10 $100 Cost of the full run, log scale Claude Opus 4.7, Open book GPT-5.4 Mini, Open book

Source: Official OGS-Bench leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-5.5 Open book 86.8 163
Claude Opus 4.7 Open book 85.9 $16 92
GPT-5.4 Open book 85.3 $1.01 67
Gemini 2.5 Pro Jun 2025 Open book 85.0 $1.03 62
GPT-4o Mar 2025 Open book 80.9 $1.47 48
Gemini 2.5 Flash Sep 2025 Open book 80.8 $0.25 65
GPT-4.1 Open book 80.5 $1.27 66
Claude Sonnet 4.6 Open book 80.3 $2.73 110
GPT-5.4 Mini Open book 79.6 $0.18 50
Claude Haiku 4.5 Open book 78.7 $0.66 82
Grok 4 Fast Open book 76.4 $0.13 59
GPT-4.1 mini Open book 76.4 $0.24 54
Gemini 2.5 Flash-Lite Sep 2025 Open book 75.1 $0.06 54
GPT-4o mini Open book 72.8 $0.08 35
GPT-5.4 Nano Open book 71.3 $0.04 80
GPT-3.5 Turbo Jan 2024 Open book 62.1 $0.26 28
GPT-4.1 nano Open book 61.9 $0.06 40

3 models have held number one on the official leaderboard since its release in April 2026

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

OpenAI Anthropic
Own leaderboard Not tested since Apr 2026
GPT-5.5 86.8

Source: Official OGS-Bench 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
60.0 70.0 80.0 90.0 2024 2025 2026 60.0 70.0 80.0 90.0 2024 2025 2026

Source: Official OGS-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 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

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 7 months ago, in April 2026, for GPT-5.5. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (20 entries).
  • Pass Models scored 17 models we track, plus 3 other entries on its leaderboard.
  • Warning Newest score The newest came 7 months ago, in April 2026, for GPT-5.5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.5 (#21).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on opengastronomy.org: 17 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