# Scale Labs SEAL Coding (SEAL Coding)

> Scale’s SEAL coding board: expert reviewers compared model answers to 1,000 prompts across programming languages, debugging, optimization, and documentation tasks. BenchLM shows it as display-only reference data, excluded from rankings. Scale marked this board deprecated as of March 2025 and no longer adds models.

Canonical page: https://benchlm.ai/benchmarks/scale-coding

- Category: [Coding](/coding)
- Last updated: November 12, 2026 snapshot

## About SEAL Coding

- Year: 2026
- Tasks: 1,000 coding prompts
- Format: Elo-style rating
- Difficulty: External agent and model evaluation
- Paper: [Scale Labs leaderboard](https://labs.scale.com/leaderboard/coding)

Reviewers checked accuracy, coherence, and relevance, with code execution tests on the answers. Scores are ratings on an Elo-style scale built from expert human comparisons, so only the gaps between rows carry meaning. BenchLM mirrors 28 published rows captured on November 12, 2026 snapshot.

SEAL Coding is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (28 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [o1-mini](/models/o1-mini) | OpenAI | 1237 |
| 2 | [o3-mini](/models/o3-mini) | OpenAI | 1137 |
| 3 | [GPT-4o (November 2024)](https://labs.scale.com/leaderboard/coding) | OpenAI | 1132 |
| 4 | [o1-preview](/models/o1-preview) | OpenAI | 1123 |
| 5 | [Gemini 2.0 Flash Experimental (December 2024)](https://labs.scale.com/leaderboard/coding) | Google | 1111 |
| 6 | [Gemini 2.0 Pro (December 2024)](https://labs.scale.com/leaderboard/coding) | Google | 1109 |
| 7 | [Gemini 2.0 Flash Thinking (January 2025)](https://labs.scale.com/leaderboard/coding) | Google | 1108 |
| 8 | [DeepSeek-R1](/models/deepseek-r1) | DeepSeek | 1100 |
| 9 | [o1 (December 2024)](https://labs.scale.com/leaderboard/coding) | OpenAI | 1083 |
| 10 | [Claude 3.5 Sonnet (June 2024)](https://labs.scale.com/leaderboard/coding) | Anthropic | 1079 |
| 11 | [GPT-4o (August 2024)](https://labs.scale.com/leaderboard/coding) | OpenAI | 1045 |
| 12 | [GPT-4o (May 2024)](https://labs.scale.com/leaderboard/coding) | OpenAI | 1036 |
| 13 | [GPT-4 Turbo Preview](https://labs.scale.com/leaderboard/coding) | OpenAI | 1034 |
| 14 | [Mistral Large 2](/models/mistral-large-2) | Mistral | 1029 |
| 15 | [Llama 3.1 405B Instruct](https://labs.scale.com/leaderboard/coding) | Meta | 1022 |
| 16 | [Gemini 1.5 Pro (August 27, 2024)](https://labs.scale.com/leaderboard/coding) | Google | 1007 |
| 17 | [Gemini 1.5 Pro (May 2024)](https://labs.scale.com/leaderboard/coding) | Google | 994 |
| 18 | [GPT-4 (November 2024)](https://labs.scale.com/leaderboard/coding) | OpenAI | 992 |
| 19 | [DeepSeek V3](/models/deepseek-v3) | DeepSeek | 985 |
| 20 | [Llama 3.2 90B Vision Instruct](https://labs.scale.com/leaderboard/coding) | Meta | 984 |
| 21 | [Claude 3 Opus](/models/claude-3-opus) | Anthropic | 959 |
| 22 | [Gemini 1.5 Flash](/models/gemini-1-5-flash) | Google | 943 |
| 23 | [Gemini 1.5 Pro (April 2024)](https://labs.scale.com/leaderboard/coding) | Google | 891 |
| 24 | [Claude 3 Sonnet](https://labs.scale.com/leaderboard/coding) | Anthropic | 879 |
| 25 | [Llama 3 70B Instruct](https://labs.scale.com/leaderboard/coding) | Meta | 871 |
| 26 | [Mistral Large](https://labs.scale.com/leaderboard/coding) | Mistral | 811 |
| 27 | [Gemini 1.0 Pro](/models/gemini-1-0-pro) | Google | 685 |
| 28 | [CodeLlama 34B Instruct](https://labs.scale.com/leaderboard/coding) | Meta | 598 |

## FAQ

### What does SEAL Coding measure?

Scale’s SEAL coding board: expert reviewers compared model answers to 1,000 prompts across programming languages, debugging, optimization, and documentation tasks. BenchLM shows it as display-only reference data, excluded from rankings. Scale marked this board deprecated as of March 2025 and no longer adds models.

### Which model leads the published SEAL Coding snapshot?

o1-mini currently leads the published SEAL Coding snapshot with a score of 1237.

### How many models are evaluated on SEAL Coding?

The November 12, 2026 snapshot contains 28 AI models.

### Does SEAL Coding affect BenchLM's overall score?

Not directly. SEAL Coding is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.