Scale Labs TutorBench (TutorBench)
We show this table for reference; we do not rank on it.
A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.
Scale score on TutorBench — November 12, 2026 snapshot
We mirror the published scale score view for TutorBench. Muse Spark leads the public snapshot at 68.5%, followed by GPT-5.4 Pro (56.6%) and gemini-2.5-pro-preview-06-05 (55.6%). We do not use these results to rank models overall.
Muse Spark
Meta
GPT-5.4 Pro
OpenAI
gemini-2.5-pro-preview-06-05
27 modelsKnowledgeCurrentDisplay onlyUpdated November 12, 2026 snapshot
Scale score table (27 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Muse SparkMeta | 68.5% | Not reported | Closed |
| 2 | GPT-5.4 ProOpenAI | 56.6% | Not reported | Closed |
| 3 | gemini-2.5-pro-preview-06-05Google | 55.6% | Not reported | Unknown |
| 4 | gpt-5-2025-08-07OpenAI | 55.3% | Not reported | Unknown |
| 5 | o3-proOpenAI | 54.6% | Not reported | Closed |
| 6 | Kimi K2.5Moonshot AI | 54.6% | Not reported | Open |
| 7 | gpt-5.1-thinkingOpenAI | 54.1% | Not reported | Unknown |
| 8 | claude-opus-4-6-thinking-maxAnthropic | 53.7% | Not reported | Unknown |
| 9 | gemini-3-pro-previewGoogle | 53.7% | Not reported | Unknown |
| 10 | claude-opus-4-6 (Non-Thinking)Anthropic | 53.5% | Not reported | Unknown |
| 11 | GPT-5.2OpenAI | 53.5% | Not reported | Closed |
| 12 | gemini-3.1-pro-previewGoogle | 53.0% | Not reported | Unknown |
| 13 | o3-2025-04-16-mediumOpenAI | 52.8% | Not reported | Unknown |
| 14 | o3-2025-04-16-highOpenAI | 52.1% | Not reported | Unknown |
| 15 | gemini-3.1-flash-lite-previewGoogle | 51.5% | Not reported | Unknown |
| 16 | Claude Opus 4.5 ThinkingAnthropic | 51.2% | Not reported | Closed |
| 17 | claude-opus-4-1-20250805-thinkingAnthropic | 50.8% | Not reported | Unknown |
| 18 | Claude Opus 4.5Anthropic | 49.8% | Not reported | Closed |
| 19 | Claude 4 Opus ThinkingAnthropic | 49.7% | Not reported | Closed |
| 20 | gpt-5.1-instantOpenAI | 49.1% | Not reported | Unknown |
| 21 | Claude Sonnet 4.5 ThinkingAnthropic | 49% | Not reported | Closed |
| 22 | claude-opus-4-1-20250805_anthropicAnthropic | 47.4% | Not reported | Unknown |
| 23 | claude-37-sonnet-thinkingAnthropic | 46.5% | Not reported | Unknown |
| 24 | Claude Sonnet 4.5Anthropic | 45.7% | Not reported | Closed |
| 25 | claude-opus-4-20250514Anthropic | 45.5% | Not reported | Unknown |
| 26 | llama4-maverickMeta | 40.2% | Not reported | Unknown |
| 27 | GPT-4oOpenAI | 36.1% | Not reported | Closed |
How to read this leaderboard
Compare the published configurations as complete evaluation systems. The source can combine a base model, agent scaffold, tools, budget, and inference setting in each result.
Operator receipt: 27 sourced rows are currently displayable on this page; the leading published row is Muse Spark at 68.5%.
Honest limit: This Scale table is display-only context, not benchmark provenance or a weighted model-only comparison.
How BenchLM shows TutorBench
BenchLM mirrors 27 published rows from Scale Labs’ public TutorBench leaderboard, captured on November 12, 2026 snapshot.
The table is display only. It is useful context for a published agent or model configuration, but it does not enter BenchLM’s overall or category rankings. Scale has not added a row to this board since April 2026.
The published TutorBench snapshot places Muse Spark first at 68.5%. The third row is 12.9 points behind. The broader top-10 range is 15.0 points, so the table still separates the published systems.
27 models have been evaluated on TutorBench. The benchmark falls in the Knowledge category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. TutorBench is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About TutorBench
Year
2026
Tasks
27 published rows
Format
Published Scale leaderboard score
Difficulty
External agent and model evaluation
BenchLM mirrors 27 published rows from the TutorBench public table captured on November 12, 2026 snapshot.
Freshness and provenance
Version
TutorBench 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does TutorBench measure?
A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.
Which model leads the published TutorBench snapshot?
Muse Spark currently leads the published TutorBench snapshot with 68.5% scale score. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on TutorBench?
The November 12, 2026 snapshot snapshot contains 27 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.