Benchmark Almanac

CSTutorBench

Gemma 4 31B IT 0.89 leads since July 2026
Not in model cards

Models tutor students who use block-based programming to control a virtual robot. A rubric grades the feedback on 17 scenarios, including debugging and answer disclosure.

Released
July 2026
Kind
Other professions
Level
University
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Occupations
Educational instruction and library
Strongest model tested
Gemma 4 31B IT #101 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
11
Entries on its own leaderboard, 8 models we track

Gemma 4 31B IT leads with 0.89 among models released by April 2026 Gemma 4 31B IT leads with 0.89 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 Gemma 4 31B IT Trial 2 system prompt 0.89
  2. 2 Qwen3.5-9B Trial 2 system prompt 0.79
  3. 3 qwen3.6_latest Trial 2 system prompt 0.78
  4. 4 deepseek-r1_8b Trial 2 system prompt 0.77
  5. 5 Nemotron 3 Super Trial 2 system prompt 0.73
  6. 6 Gemma 4 E4B Trial 2 system prompt 0.72
  7. 7 Gemma 3 27B Trial 2 system prompt 0.69
  8. 8 gpt-oss-20b Trial 2 system prompt 0.64
  9. 9 Qwen3-Coder-30B-A3B-Instruct Trial 1 system prompt 0.53
  10. 10 Gemma 3 4B Trial 2 system prompt 0.50
  11. 11 olmo-3_7b-think Trial 2 system prompt 0.49

Source: Official CSTutorBench leaderboard

#EntryScoreSettingDateSource
1 Gemma 4 31B IT 0.89 Trial 2 system prompt Apr 2026 Official leaderboard
2 Qwen3.5-9B 0.79 Trial 2 system prompt Feb 2026 Official leaderboard
3 qwen3.6_latest 0.78 Trial 2 system prompt Official leaderboard
4 deepseek-r1_8b 0.77 Trial 2 system prompt Official leaderboard
5 Nemotron 3 Super 0.73 Trial 2 system prompt Mar 2026 Official leaderboard
6 Gemma 4 E4B 0.72 Trial 2 system prompt Mar 2026 Official leaderboard
7 Gemma 3 27B 0.69 Trial 2 system prompt Mar 2025 Official leaderboard
8 gpt-oss-20b 0.64 Trial 2 system prompt Aug 2025 Official leaderboard
9 Qwen3-Coder-30B-A3B-Instruct 0.53 Trial 1 system prompt Jul 2025 Official leaderboard
10 Gemma 3 4B 0.50 Trial 2 system prompt Mar 2025 Official leaderboard
11 olmo-3_7b-think 0.49 Trial 2 system prompt Official leaderboard
#ModelScoreSettingDateSource
1 Gemma 4 31B IT 0.89 Trial 2 system prompt Apr 2026 Official leaderboard
2 Qwen3.5-9B 0.79 Trial 2 system prompt Feb 2026 Official leaderboard
3 Nemotron 3 Super 0.73 Trial 2 system prompt Mar 2026 Official leaderboard
4 Gemma 4 E4B 0.72 Trial 2 system prompt Mar 2026 Official leaderboard
5 Gemma 3 27B 0.69 Trial 2 system prompt Mar 2025 Official leaderboard
6 gpt-oss-20b 0.64 Trial 2 system prompt Aug 2025 Official leaderboard
7 Qwen3-Coder-30B-A3B-Instruct 0.53 Trial 1 system prompt Jul 2025 Official leaderboard
8 Gemma 3 4B 0.50 Trial 2 system prompt Mar 2025 Official leaderboard

Rankings by part

We read 11 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Conciseness, Vocabulary and Accuracy. Gemma 4 31B IT leads 4 of them, and 4 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Conciseness

8 models we track have a score here.

  1. 1 Gemma 4 31B IT 1.00
  2. 2 Gemma 4 E4B ↑2 0.65
  3. 3 Qwen3.5-9B ↓1 0.62
  4. 4 gpt-oss-20b ↑2 0.41
  5. 5 Nemotron 3 Super ↓2 0.32
  6. 6 Gemma 3 27B ↓1 0.26
  7. 7 Qwen3-Coder-30B-A3B-Instruct 0.21
  8. 8 Gemma 3 4B 0.09

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Gemma 4 31B IT
Trial 1 sys… Trial 2 sys…
0.89
Qwen3.5-9B
Trial 1 sys… Trial 2 sys…
0.79
Nemotron 3 Super
Trial 1 sys… Trial 2 sys…
0.73
Gemma 4 E4B
Trial 1 sys… Trial 2 sys…
0.72
Gemma 3 27B
Trial 1 sys… Trial 2 sys…
0.69
gpt-oss-20b
Trial 1 sys… Trial 2 sys…
0.64
Qwen3-Coder-30B-A3B-Instruct
Trial 2 sys…
0.53
Gemma 3 4B
Trial 1 sys… Trial 2 sys…
0.50

Source: Official CSTutorBench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemma 4 31B IT Trial 2 system prompt 0.89
Qwen3.5-9B Trial 2 system prompt 0.79
Gemma 4 31B IT Trial 1 system prompt 0.77
Nemotron 3 Super Trial 2 system prompt 0.73
Gemma 4 E4B Trial 2 system prompt 0.72
Gemma 3 27B Trial 2 system prompt 0.69
Qwen3.5-9B Trial 1 system prompt 0.66
Gemma 4 E4B Trial 1 system prompt 0.65
gpt-oss-20b Trial 2 system prompt 0.64
Gemma 3 27B Trial 1 system prompt 0.58
Nemotron 3 Super Trial 1 system prompt 0.57
gpt-oss-20b Trial 1 system prompt 0.55
Qwen3-Coder-30B-A3B-Instruct Trial 1 system prompt 0.53
Qwen3-Coder-30B-A3B-Instruct Trial 2 system prompt 0.52
Gemma 3 4B Trial 2 system prompt 0.50
Gemma 3 4B Trial 1 system prompt 0.36

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Gemma 4 31B IT
Trial 1 sys… Trial 2 sys…
0.89
Qwen3.5-9B
Trial 1 sys… Trial 2 sys…
0.79
Nemotron 3 Super
Trial 1 sys… Trial 2 sys…
0.73
Gemma 4 E4B
Trial 1 sys… Trial 2 sys…
0.72
Gemma 3 27B
Trial 1 sys… Trial 2 sys…
0.69
gpt-oss-20b
Trial 1 sys… Trial 2 sys…
0.64
Qwen3-Coder-30B-A3B-Instruct
Trial 2 sys…
0.53
Gemma 3 4B
Trial 1 sys… Trial 2 sys…
0.50

Source: Official CSTutorBench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemma 4 31B IT Trial 2 system prompt 0.89
Qwen3.5-9B Trial 2 system prompt 0.79
Gemma 4 31B IT Trial 1 system prompt 0.77
Nemotron 3 Super Trial 2 system prompt 0.73
Gemma 4 E4B Trial 2 system prompt 0.72
Gemma 3 27B Trial 2 system prompt 0.69
Qwen3.5-9B Trial 1 system prompt 0.66
Gemma 4 E4B Trial 1 system prompt 0.65
gpt-oss-20b Trial 2 system prompt 0.64
Gemma 3 27B Trial 1 system prompt 0.58
Nemotron 3 Super Trial 1 system prompt 0.57
gpt-oss-20b Trial 1 system prompt 0.55
Qwen3-Coder-30B-A3B-Instruct Trial 1 system prompt 0.53
Qwen3-Coder-30B-A3B-Instruct Trial 2 system prompt 0.52
Gemma 3 4B Trial 2 system prompt 0.50
Gemma 3 4B Trial 1 system prompt 0.36

Number one over time

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

Google DeepMind
Own leaderboard Not tested since Jul 2026

Source: Official CSTutorBench 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
0.40 0.60 0.80 1.00 2025 2026 0.40 0.60 0.80 1.00 2025 2026

Source: Official CSTutorBench 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 7 months ago, in April 2026, for Gemma 4 31B 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 official leaderboard (11 entries).
  • Pass Models scored 8 models we track, plus 3 other entries on its leaderboard.
  • Warning Newest score The newest came 7 months ago, in April 2026, for Gemma 4 31B IT.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Gemma 4 31B IT (#101).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 16 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