Scale Labs SEAL Chinese (SEAL Chinese)
We show this table for reference; we do not rank on it.
Scale's SEAL leaderboard for Chinese: expert raters judge how well models answer about 1,000 prompts written by native Chinese speakers. 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.
Rating on SEAL Chinese — November 12, 2026 snapshot
We mirror the published rating view for SEAL Chinese. o1 (December 2024) leads the public snapshot at 1165, followed by o3-mini (1156) and Gemini 1.5 Pro (August 27, 2024) (1120). We do not use these results to rank models overall.
o1 (December 2024)
OpenAI
o3-mini
OpenAI
Gemini 1.5 Pro (August 27, 2024)
21 modelsCross-domainCurrentDisplay onlyUpdated November 12, 2026 snapshot
Rating table (21 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | o1 (December 2024)OpenAI | 1165 | Not reported | Unknown |
| 2 | o3-miniOpenAI | 1156 | Not reported | Closed |
| 3 | Gemini 1.5 Pro (August 27, 2024)Google | 1120 | Not reported | Unknown |
| 4 | o1-previewOpenAI | 1120 | Not reported | Closed |
| 5 | Gemini 2.0 Pro (December 2024)Google | 1117 | Not reported | Unknown |
| 6 | Gemini Pro Flash 2Google | 1115 | Not reported | Unknown |
| 7 | Gemini 1.5 Pro (November 2024)Google | 1077 | Not reported | Unknown |
| 8 | Gemini 2.0 Flash Thinking (January 2025)Google | 1060 | Not reported | Unknown |
| 9 | DeepSeek-R1DeepSeek | 1052 | Not reported | Open |
| 10 | DeepSeek V3DeepSeek | 1031 | Not reported | Open |
| 11 | GPT-4o (August 2024)OpenAI | 1029 | Not reported | Unknown |
| 12 | Gemini 1.5 FlashGoogle | 1015 | Not reported | Closed |
| 13 | Mistral Large 2Mistral | 1006 | Not reported | Closed |
| 14 | DeepSeek V2 ChatDeepseek | 996 | Not reported | Unknown |
| 15 | GPT-4 (November 2024)OpenAI | 985 | Not reported | Unknown |
| 16 | Aya Expanse 32BCohere | 967 | Not reported | Unknown |
| 17 | Gemma 2 27BGoogle | 966 | Not reported | Open |
| 18 | Claude 3.5 Sonnet (June 2024)Anthropic | 930 | Not reported | Unknown |
| 19 | Qwen 2 72B InstructAlibaba | 902 | Not reported | Unknown |
| 20 | Llama 3.3 70B InstructMeta | 883 | Not reported | Unknown |
| 21 | Yi 1.5 34B ChatUnknown | 780 | 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: 21 sourced rows are currently displayable on this page; the leading published row is o1 (December 2024) at 1165.
Honest limit: This Scale table is display-only context, not benchmark provenance or a weighted model-only comparison. Scale marked this board deprecated as of March 2025 and no longer adds models.
How BenchLM shows SEAL Chinese
BenchLM mirrors 21 published rows from Scale Labs’ public SEAL Chinese 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 marked this board deprecated as of March 2025 and no longer adds models.
Snapshot
The published SEAL Chinese snapshot places o1 (December 2024) first at 1165. The third row is 45 score units behind. The broader top-10 range is 134 score units, so the table still separates the published systems.
21 models have been evaluated on SEAL Chinese. The benchmark falls in the Cross-domain category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. SEAL Chinese is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About SEAL Chinese
Year
2026
Tasks
About 1,000 prompts per language
Format
Elo-style rating
Difficulty
External agent and model evaluation
Single-turn prompts across general and culturally specific scenarios, rated by native-speaking experts. Scores are ratings on an Elo-style scale built from expert human comparisons, so only the gaps between rows carry meaning. BenchLM mirrors 21 published rows captured on November 12, 2026 snapshot.
Freshness and provenance
Version
SEAL Chinese 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 SEAL Chinese measure?
Scale's SEAL leaderboard for Chinese: expert raters judge how well models answer about 1,000 prompts written by native Chinese speakers. 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 Chinese snapshot?
o1 (December 2024) currently leads the published SEAL Chinese snapshot with 1165 rating. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on SEAL Chinese?
The November 12, 2026 snapshot snapshot contains 21 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.