Benchmark Almanac

BEA 2023 Shared Task on Generating AI Teacher Responses in Educational Dialogues

NAISTeacher 1.50 tops its leaderboard GPT-4 Jun 2023 3.00 leads the models we track
Not in model cards

Models write the teacher's next reply in an English lesson. Organizers average the system's ranks on BERTScore F1 and DialogRPT.

Eight teams participated. Human raters also compared the three highest ranked submissions on how well each reply helped and understood the student.

Released
March 2023
Kind
Multi-turn conversation
Level
University
Scored
Graded by a model
Answers
Open-ended
Tasks
Public, plus a private test set
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
A host
Occupations
Educational instruction and library
Strongest model tested
GPT-4 Jun 2023 #186 on the capability index
Last model added
June 2023
Scores refreshed
November 20th, 2026
9
Entries on its own leaderboard, 1 model we track

NAISTeacher tops the leaderboard with 1.50 GPT-4 leads with 3.00

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. The best model we track is GPT-4, at 3.00. This view shows only models we track, each at its best.

  1. 1 NAISTeacher Submission 1 1.50
  2. 2 adaio Submission 2 3.00
  3. 3 GPT-4 Jun 2023, Submission 3 3.00
  4. 4 opt2.7-1LineEach Submission 5 4.50
  5. 5 teacher Submission 4 4.50
  6. 6 bea Task UNTRAINED BEA BERTScore Only Only Response Submission 7 6.00
  7. 7 Retuyt-InCo-AlpacaEmb Submission 8 7.50
  8. 8 distilgpt2 Submission 9 9.50
  9. 9 TanTanLabs-zero-shot-with-filler Submission 10 9.50

Source: Official BEA 2023 Shared Task on Generating AI Teacher Responses in Educational Dialogues leaderboard

Source: Official BEA 2023 Shared Task on Generating AI Teacher Responses in Educational Dialogues leaderboard

#EntryScoreSettingSubmitted byDateSource
1 NAISTeacher 1.50 Submission 1 NAIST Official leaderboard
2 adaio 3.00 Submission 2 NBU Official leaderboard
3 GPT-4 Jun 2023 3.00 Submission 3 Cornell Jun 2023 Official leaderboard
4 opt2.7-1LineEach 4.50 Submission 5 Retuyt-InCo Official leaderboard
5 teacher 4.50 Submission 4 aiitis Official leaderboard
6 bea Task UNTRAINED BEA BERTScore Only Only Response 6.00 Submission 7 Data Science - NLP - HSG Official leaderboard
7 Retuyt-InCo-AlpacaEmb 7.50 Submission 8 Retuyt-InCo Official leaderboard
8 distilgpt2 9.50 Submission 9 DT Official leaderboard
9 TanTanLabs-zero-shot-with-filler 9.50 Submission 10 TanTanLabs Official leaderboard
#ModelScoreSettingSubmitted byDateSource
1 GPT-4 Jun 2023 3.00 Submission 3 Cornell Jun 2023 Official leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
3.00 4.00 5.00 6.00 2023 3.00 4.00 5.00 6.00 2023

Source: Official BEA 2023 Shared Task on Generating AI Teacher Responses in Educational Dialogues 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 Instructions and chat

All 144 in Instructions and chat
  • IFEval In use Models answer about 500 prompts containing verifiable instructions such as word limits or required keywords. Code checks each instruction for strict and loose accuracy. Gemma 4 31B IT 98.9%
  • IFBench In use 58 new out-of-domain verifiable output constraints. Scored by programmatic constraint checks (strict or loose). Flint 89.2%
  • MultiChallenge In use Multi-turn conversations testing instruction retention, inference memory, versioned editing and self-coherence. Graded by LLM judge with rubrics. GPT-6.1 Sol 82%
  • Multi-IF In use Models follow verifiable instructions over three-turn conversations in eight languages, extending IFEval. Scored by instruction accuracy at each turn. LFM3-1.7B 89.1%
  • AdvancedIF In use Models respond to over 1,600 prompts with complex, multi-turn, and system-level instructions. A judge model checks each response against expert-written rubrics. MAI-Thinking-1 85%
  • MM-IFEval In use Multimodal models follow image-grounded instructions with exact output constraints. Checks mix rule-based verification and a judge model. GPT-4o 64.6%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 3 years ago, in June 2023, for GPT-4. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (9 entries).
  • Pass Models scored 1 model we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 3 years ago, in June 2023, for GPT-4.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-4 (#186).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on sig-edu.org: 2 rows for models we track and 8 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