Benchmark Almanac

MisconceptDialogue

Claude Haiku 4.5 87.8% leads since October 2026
Not in model cards

Tutors identify a simulated student's mathematical misconception and address it through dialogue. The authors compare exact-match and human-calibrated model grading on the same conversations.

Released
October 2026
Kind
Other professions
Level
University
Scored
Graded by a model
Answers
Agent task
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Multi-turn
Items from
Generated
Refresh
One fixed set
Run by
A host
Occupations
Educational instruction and library
Strongest model tested
Gemini 3.1 Flash-Lite #96 on the capability index
Last model added
March 2026
Scores refreshed
November 20th, 2026
7
Entries in its paper, 4 models we track

Claude Haiku 4.5 leads with 87.8% among models released by March 2026 Claude Haiku 4.5 leads with 87.8% among models released by March 2026

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The score is the percentage of simulated tutoring dialogues that GPT-5-mini judged successful. 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 March 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 Claude Haiku 4.5 87.8%
  2. 2 Gemini 3.1 Flash-Lite 86.2%
  3. 3 Qwen3.5-9B 84.4%
  4. 4 Phi-4 83.6%
  5. 5 Qwen2.5-7B-Instruct 81.8%
  6. 6 Llama-3.1-8B-Instruct 79.1%
  7. 7 Mistral-7B-Instruct-v0.3 77.8%

Source: The MisconceptDialogue paper

#EntryScoreSettingDateSource
1 Claude Haiku 4.5 87.8% Oct 2025 Official leaderboard
2 Gemini 3.1 Flash-Lite 86.2% Mar 2026 Official leaderboard
3 Qwen3.5-9B 84.4% Feb 2026 Official leaderboard
4 Phi-4 83.6% Dec 2024 Official leaderboard
5 Qwen2.5-7B-Instruct 81.8% Official leaderboard
6 Llama-3.1-8B-Instruct 79.1% Official leaderboard
7 Mistral-7B-Instruct-v0.3 77.8% Official leaderboard
#ModelScoreSettingDateSource
1 Claude Haiku 4.5 87.8% Oct 2025 Official leaderboard
2 Gemini 3.1 Flash-Lite 86.2% Mar 2026 Official leaderboard
3 Qwen3.5-9B 84.4% Feb 2026 Official leaderboard
4 Phi-4 83.6% Dec 2024 Official leaderboard

Rankings by part

We read 4 parts from the benchmark’s paper, each ranking models on one part of the test, such as Strict match, GPT-5-mini judge and Claude Haiku 4.5 judge. Claude Haiku 4.5 leads 3 of them, and Qwen3.5-9B leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Strict match

4 models we track have a score here.

  1. 1 Qwen3.5-9B ↑2 43.3%
  2. 2 Claude Haiku 4.5 ↓1 43.1%
  3. 3 Gemini 3.1 Flash-Lite ↓1 43.1%
  4. 4 Phi-4 42.4%

Number one over time

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

Anthropic
Benchmark paper Passed 85% in Oct 2026, by Claude Haiku 4.5

Source: The MisconceptDialogue paper

Claude Haiku 4.5 on the official leaderboard passed 85% in October 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
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Haiku 4.5 Phi-4 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Haiku 4.5

Source: Official MisconceptDialogue 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

7 of 8 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the benchmark’s paper (7 entries).
  • Pass Models scored 4 models we track, plus 3 other entries on its leaderboard.
  • Note Newest score The newest came 8 months ago, in March 2026, for Gemini 3.1 Flash-Lite. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Gemini 3.1 Flash-Lite (#96).
  • Note Headroom Saturated. The best is 87.8%, past 85% since October 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 4 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