Scale Labs MASK (MASK)
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 MASK — November 12, 2026 snapshot
We mirror the published scale score view for MASK. claude-opus-4-6 (Non-Thinking) leads the public snapshot at 96.3%, followed by Claude Sonnet 4.5 Thinking (96.1%) and Claude Sonnet 4 (Thinking) (95.3%). We do not use these results to rank models overall.
claude-opus-4-6 (Non-Thinking)
Anthropic
Claude Sonnet 4.5 Thinking
Anthropic
Claude Sonnet 4 (Thinking)
Anthropic
67 modelsInstruction FollowingCurrentDisplay onlyUpdated November 12, 2026 snapshot
Scale score table (67 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | claude-opus-4-6 (Non-Thinking)Anthropic | 96.3% | Not reported | Unknown |
| 2 | Claude Sonnet 4.5 ThinkingAnthropic | 96.1% | Not reported | Closed |
| 3 | Claude Sonnet 4 (Thinking)Anthropic | 95.3% | Not reported | Unknown |
| 4 | claude-opus-4-1-20250805-thinkingAnthropic | 94.2% | Not reported | Unknown |
| 5 | Claude Opus 4.5 ThinkingAnthropic | 92.5% | Not reported | Closed |
| 6 | GPT-OSS 120BOpenAI | 92% | Not reported | Open |
| 7 | GPT-5.4 ProOpenAI | 91.7% | Not reported | Closed |
| 8 | gpt-5.4-2026-03-05 (xhigh thinking)OpenAI | 89.7% | Not reported | Unknown |
| 9 | Claude Sonnet 4Anthropic | 89.3% | Not reported | Unknown |
| 10 | Claude Opus 4 (Thinking)Anthropic | 87.9% | Not reported | Unknown |
| 11 | claude-opus-4-1-20250805Anthropic | 87.4% | Not reported | Unknown |
| 12 | Claude Opus 4.5Anthropic | 87.1% | Not reported | Closed |
| 13 | GPT-5.2OpenAI | 86.7% | Not reported | Closed |
| 14 | GPT-OSS 20BOpenAI | 86.5% | Not reported | Open |
| 15 | Claude Sonnet 4.5Anthropic | 86.4% | Not reported | Closed |
| 16 | gpt-5.1-thinkingOpenAI | 86.3% | Not reported | Unknown |
| 17 | gpt-5-pro-2025-10-06OpenAI | 86.0% | Not reported | Unknown |
| 18 | claude-opus-4-6-thinking-maxAnthropic | 85.4% | Not reported | Unknown |
| 19 | o3 (high) (April 2025)OpenAI | 84.5% | Not reported | Unknown |
| 20 | GPT-5 miniOpenAI | 82.6% | Not reported | Closed |
| 21 | o3 (medium) (April 2025)OpenAI | 82.6% | Not reported | Unknown |
| 22 | o3 Pro (high) (June 2025)OpenAI | 82.5% | Not reported | Unknown |
| 23 | Claude 3.7 Sonnet (Thinking) (February 2025)Anthropic | 82.1% | Not reported | Unknown |
| 24 | Claude Opus 4Anthropic | 80.3% | Not reported | Unknown |
| 25 | gpt-5-2025-08-07OpenAI | 79.3% | Not reported | Unknown |
| 26 | Claude 3 OpusAnthropic | 79% | Not reported | Closed |
| 27 | o4-mini (high) (April 2025)OpenAI | 78.6% | Not reported | Unknown |
| 28 | o4-mini (medium) (April 2025)OpenAI | 72.9% | Not reported | Unknown |
| 29 | Claude 3.5 Sonnet (October 2024)Anthropic | 72.3% | Not reported | Unknown |
| 30 | Claude 3.7 Sonnet (February 2025)Anthropic | 72.3% | Not reported | Unknown |
| 31 | Kimi K2.5Moonshot AI | 70.5% | Not reported | Open |
| 32 | gpt-5.1-instantOpenAI | 63.3% | Not reported | Unknown |
| 33 | o1-proOpenAI | 61.6% | Not reported | Closed |
| 34 | glm-4p5Z.AI | 61.5% | Not reported | Unknown |
| 35 | GPT-4.1 nano\nOpenAI | 61.4% | Not reported | Unknown |
| 36 | Llama 3.1 405B InstructMeta | 61.4% | Not reported | Unknown |
| 37 | glm-4p5-airZ.AI | 60.8% | Not reported | Unknown |
| 38 | gpt 4o (November 2024)OpenAI | 60.1% | Not reported | Unknown |
| 39 | o1 (December 2024)OpenAI | 59.3% | Not reported | Unknown |
| 40 | Deepseek R1 (Jan 2025)Deepseek | 57.3% | Not reported | Unknown |
| 41 | GPT 4.5 PreviewOpenAI | 56.9% | Not reported | Unknown |
| 42 | Mistral MagistralMistral | 56.5% | Not reported | Unknown |
| 43 | Qwen3 235B-A22BAlibaba | 56.4% | Not reported | Open |
| 44 | Gemini 2.5 Pro Experimental (March 2025)Google | 55.9% | Not reported | Unknown |
| 45 | gemini-2.5-pro-preview-06-05Google | 55.7% | Not reported | Unknown |
| 46 | Llama 3.2 90B Vision InstructMeta | 54.1% | Not reported | Unknown |
| 47 | Gemini 2.5 Pro Preview (May 06 2025)Google | 53.1% | Not reported | Unknown |
| 48 | DeepSeek-R1-0528DeepSeek | 53% | Not reported | Open |
| 49 | Llama 3.3 70B InstructMeta | 51.9% | Not reported | Unknown |
| 50 | GPT-4.1OpenAI | 51.1% | Not reported | Closed |
| 51 | GPT-4.1 miniOpenAI | 50% | Not reported | Closed |
| 52 | Llama 4 MaverickMeta | 49.7% | Not reported | Open |
| 53 | o3 mini (Low)OpenAI | 49.7% | Not reported | Unknown |
| 54 | Gemini 2.0 Flash Thinking (January 2025)Google | 49.5% | Not reported | Unknown |
| 55 | Gemini 2.5 Flash Preview (May 2025)Google | 49.1% | Not reported | Unknown |
| 56 | Gemini 2.0 FlashGoogle | 49.1% | Not reported | Closed |
| 57 | o3 mini (Medium)OpenAI | 48.9% | Not reported | Unknown |
| 58 | Gemini 2.0 Pro Experimental (February 2025)Google | 48.7% | Not reported | Unknown |
| 59 | gemini-3.1-flash-lite-previewGoogle | 48.4% | Not reported | Unknown |
| 60 | Mistral Large 2411Mistral | 47.5% | Not reported | Unknown |
| 61 | o3 mini (High)OpenAI | 46.8% | Not reported | Unknown |
| 62 | kimi-k2-instructMoonshot AI | 46.7% | Not reported | Unknown |
| 63 | DeepSeek V3.1DeepSeek | 46.3% | Not reported | Open |
| 64 | Deepseek V3 (March 2025)Deepseek | 44.5% | Not reported | Unknown |
| 65 | Mistral Medium 3Mistral | 42.6% | Not reported | Closed |
| 66 | gemini-3-pro-previewGoogle | 42.6% | Not reported | Unknown |
| 67 | gemini-3.1-pro-previewGoogle | 42.4% | Not reported | Unknown |
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: 67 sourced rows are currently displayable on this page; the leading published row is claude-opus-4-6 (Non-Thinking) at 96.3%.
Honest limit: This Scale table is display-only context, not benchmark provenance or a weighted model-only comparison.
How BenchLM shows MASK
BenchLM mirrors 67 published rows from Scale Labs’ public MASK 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 March 2026.
The published MASK snapshot places claude-opus-4-6 (Non-Thinking) first at 96.3%. The third row is 1.0 points behind. The broader top-10 range is 8.4 points, so many of the published results sit in a relatively narrow band.
67 models have been evaluated on MASK. The benchmark falls in the Instruction Following category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. MASK is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About MASK
Year
2026
Tasks
67 published rows
Format
Published Scale leaderboard score
Difficulty
External agent and model evaluation
BenchLM mirrors 67 published rows from the MASK public table captured on November 12, 2026 snapshot.
Freshness and provenance
Version
MASK 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 MASK measure?
A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.
Which model leads the published MASK snapshot?
claude-opus-4-6 (Non-Thinking) currently leads the published MASK snapshot with 96.3% scale score. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on MASK?
The November 12, 2026 snapshot snapshot contains 67 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.