Benchmark Almanac

RiceChem

RoBERTa-large-mnli 0.89 tops its leaderboard Mistral-7B-Instruct-v0.2 0.43 leads the models we track
Not in model cards

Models grade long answers to four college chemistry questions. Each of the 1,264 answers has human judgments for the applicable rubric items.

Sonkar and colleagues report 0.689 F1 for GPT-4 on rubric items. S-GRADES evaluated three models on total scores for the same four questions.

Released
April 2024
Kind
Other professions
Level
University
Scored
Checked automatically
Answers
Short answer
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
Phi-2 #225 on the capability index
Last model added
April 2024
Scores refreshed
November 20th, 2026
19
Entries in its paper, 6 models we track

RoBERTa-large-mnli tops the leaderboard with 0.89 Mistral-7B-Instruct-v0.2 leads with 0.43 among models released by April 2024

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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. The best model we track is Mistral-7B-Instruct-v0.2, at 0.43. This view shows only models we track, each at its best.

The newest model with a score here came out in April 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 RoBERTa-large-mnli Fine-tuned 0.89
  2. 2 BART-large-mnli Fine-tuned 0.88
  3. 3 BART-large Fine-tuned 0.86
  4. 4 RoBERTa-large Fine-tuned 0.86
  5. 5 BART-base Fine-tuned 0.86
  6. 6 RoBERTa-base Fine-tuned 0.86
  7. 7 BERT-base Fine-tuned 0.85
  8. 8 BERT-large Fine-tuned 0.85
  9. 9 GPT4 (snapshot unspecified) Zero-shot 0.69
  10. 10 GPT3.5 (snapshot unspecified) Zero-shot 0.52
  11. 11 Qwen1.5 32B Chat Zero-shot 0.46
  12. 12 Mistral-7B-Instruct-v0.2 Zero-shot 0.43
  13. 13 OLMo-7B-Instruct Zero-shot 0.40
  14. 14 Gemma 1.1 7B IT Zero-shot 0.35
  15. 15 Yi 34B Chat Zero-shot 0.34
  16. 16 Zephyr 7B Beta Zero-shot 0.18
  17. 17 Qwen1.5-14B-Chat Zero-shot 0.16
  18. 18 Phi-2 Zero-shot 0.16
  19. 19 Vicuna 13B v1.5 Zero-shot 0.15

Source: The RiceChem paper

#EntryScoreSettingDateSource
1 RoBERTa-large-mnli 0.89 Fine-tuned Official leaderboard
2 BART-large-mnli 0.88 Fine-tuned Official leaderboard
3 BART-large 0.4B 0.86 Fine-tuned Official leaderboard
4 RoBERTa-large 0.4B 0.86 Fine-tuned Official leaderboard
5 BART-base 0.1B 0.86 Fine-tuned Official leaderboard
6 RoBERTa-base 0.1B 0.86 Fine-tuned Official leaderboard
7 BERT-base 0.1B 0.85 Fine-tuned Official leaderboard
8 BERT-large 0.3B 0.85 Fine-tuned Official leaderboard
9 GPT4 (snapshot unspecified) 0.69 Zero-shot Official leaderboard
10 GPT3.5 (snapshot unspecified) 0.52 Zero-shot Official leaderboard
11 Qwen1.5 32B Chat 0.46 Zero-shot Official leaderboard
12 Mistral-7B-Instruct-v0.2 0.43 Zero-shot Dec 2023 Official leaderboard
13 OLMo-7B-Instruct 0.40 Zero-shot Feb 2024 Official leaderboard
14 Gemma 1.1 7B IT 0.35 Zero-shot Apr 2024 Official leaderboard
15 Yi 34B Chat 0.34 Zero-shot Official leaderboard
16 Zephyr 7B Beta 0.18 Zero-shot Oct 2023 Official leaderboard
17 Qwen1.5-14B-Chat 0.16 Zero-shot Feb 2024 Official leaderboard
18 Phi-2 0.16 Zero-shot Dec 2023 Official leaderboard
19 Vicuna 13B v1.5 0.15 Zero-shot Official leaderboard
#ModelScoreSettingDateSource
1 Mistral-7B-Instruct-v0.2 0.43 Zero-shot Dec 2023 Official leaderboard
2 OLMo-7B-Instruct 0.40 Zero-shot Feb 2024 Official leaderboard
3 Gemma 1.1 7B IT 0.35 Zero-shot Apr 2024 Official leaderboard
4 Zephyr 7B Beta 0.18 Zero-shot Oct 2023 Official leaderboard
5 Qwen1.5-14B-Chat 0.16 Zero-shot Feb 2024 Official leaderboard
6 Phi-2 0.16 Zero-shot Dec 2023 Official leaderboard

Number one over time

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

Mistral AI
Benchmark paper Not tested since Apr 2024

Source: The RiceChem 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
0.20 0.30 0.40 0.50 2024 0.20 0.30 0.40 0.50 2024

Source: Official RiceChem 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 Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 2 years ago, in April 2024, for Gemma 1.1 7B IT. 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 benchmark’s paper (19 entries).
  • Pass Models scored 6 models we track, plus 13 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in April 2024, for Gemma 1.1 7B IT.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Phi-2 (#225).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 6 rows for models we track and 13 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