BIG-Bench Hard leaderboard
As of 6 Nov 2026, the top score on BIG-Bench Hard is 89.2% by Gemini 1.5 Pro 001 (Google), ahead of DeepSeek V3 at 87.5%, across 43 model configurations with a published result (Epoch AI Benchmarking Hub, result dated 14 May 2024). Twenty-three hard tasks from BIG-Bench. BIG-Bench Hard is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- Gemini 1.5 Pro 001 holds the top BIG-Bench Hard score at 89.2% (Epoch AI Benchmarking Hub, 14 May 2024).
- DeepSeek V3 is second on BIG-Bench Hard at 87.5% (Epoch AI Benchmarking Hub, 26 Dec 2024).
- BIG-Bench Hard: Twenty-three hard tasks from BIG-Bench. A 2022 benchmark, saturated for current models; kept for history (Epoch AI Benchmarking Hub).
- 43 model configurations have a published BIG-Bench Hard result on BenchLeader (BenchLeader, 6 Nov 2026).
- BIG-Bench Hard does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Twenty-three hard tasks from BIG-Bench. A 2022 benchmark, saturated for current models; kept for history.
- Published by
- Epoch AI Benchmarking Hub
- Category
- Reasoning
- Index weight
- Reference only
- Models
- 43
- Data as of
- 6 Nov 2026
CC BY 4.0 — Epoch AI, ‘AI Benchmarking Hub’, epoch.ai/benchmarks. Mirrored boards credit their original publishers.
What the test looks like
Twenty-three hard tasks from BIG-Bench.
How it is scored
Accuracy, as compiled by Epoch AI from published results.
What to keep in mind
Saturated: frontier models score near the ceiling, so it separates only older or smaller models, and results are largely self-reported by developers.
- 1Gemini 1.5 Pro 00189.2%
- 2DeepSeek V387.5%
- 3Gemini 1.5 Pro 001 Feb2484.0%
- 4Llama 3.1 405B82.9%
- 5Phi 3 (medium)81.4%
- 6Qwen2.5 72B79.8%
- 7Phi 3 Small79.1%
- 8Deepseek v278.8%
- 9GPT 4 061375.1%
- 10Phi 3 Mini71.7%
- 11Yi-34B71.7%
- 12Stable Beluga 269.3%
- 13Llama 2-70B64.9%
- 14GPT-3.5 Turbo (older v0613)61.6%
- 15Phi-259.4%
| # | ||||
|---|---|---|---|---|
| 1 |  Gemini 1.5 Pro 001 | 89.2% | – | 2024-05-14 |
| 2 |  DeepSeek V3open ↗ | 87.5% | 44.4 | 2024-12-26 |
| 3 | Gemini 1.5 Pro 001 Feb24 | 84.0% | – | 2024-02-15 |
| 4 |  Llama 3.1 405Bopen ↗ | 82.9% | 41.9 | 2024-07-23 |
| 5 |  Phi 3medium | 81.4% | 37.9 | 2024-04-23 |
| 6 |  Qwen2.5 72Bopen ↗ | 79.8% | 43.2 | 2024-09-19 |
| 7 | Phi 3 Small | 79.1% | – | 2024-04-23 |
| 8 | Deepseek v2open ↗ | 78.8% | – | 2024-05-07 |
| 9 |  GPT 4 0613 | 75.1% | 39.5 | 2023-06-13 |
| 10 | Phi 3 Miniopen | 71.7% | 36.9 | 2024-04-23 |
| 11 |  Yi-34B | 71.7% | 34.9 | 2023-11-22 |
| 12 |  Stable Beluga 2open | 69.3% | – | 2023-07-20 |
| 13 | Llama 2-70Bopen ↗ | 64.9% | 33.6 | 2023-07-18 |
| 14 | GPT-3.5 Turbo (older v0613) | 61.6% | – | 2023-06-13 |
| 15 | Phi-2open ↗ | 59.4% | – | 2023-12-12 |
| 16 |  Nemotron-4 15B | 58.7% | – | 2024-02-26 |
| 17 | LLaMA-65Bopen | 58.4% | – | 2023-02-24 |
| 18 | Llama 2-13Bopen ↗ | 58.2% | – | 2023-07-18 |
| 19 |  Mistral 7B v0.3open | 56.1% | 29.0 | 2023-09-27 |
| 20 | Gemma 7B | 55.1% | – | 2024-02-21 |
| 21 | Qwen-14B | 55.0% | – | 2023-09-24 |
| 22 |  InternLM 20B | 52.5% | – | 2023-09-18 |
| 23 | LLaMA-33B | 50.0% | – | 2023-02-24 |
| 24 |  Baichuan2-13Bopen | 49.0% | – | 2023-09-06 |
| 25 | Yi 6B | 47.2% | – | 2023-11-22 |
| 26 | Qwen-7B | 45.0% | – | 2023-09-28 |
| 27 | Llama 2-34B | 44.1% | – | 2023-07-18 |
| 28 |  Vicuna 13B v1.1 | 43.0% | – | 2023-04-12 |
| 29 | Baichuan 1-13B | 43.0% | – | 2023-07-11 |
| 30 | Baichuan 2 7B Base | 41.6% | – | 2023-09-20 |
| 31 | Llama 2-7Bopen ↗ | 39.2% | – | 2023-07-18 |
| 32 |  MPT-30B | 38.0% | – | 2023-06-22 |
| 33 | LLaMA-13B | 37.9% | – | 2023-02-24 |
| 34 |  Falcon-40B | 37.1% | – | 2023-03-15 |
| 35 | InternLM 7B | 37.0% | – | 2023-07-05 |
| 36 | MPT-7B | 35.6% | – | 2023-05-05 |
| 37 | Gemma 2B | 35.2% | – | 2024-02-21 |
| 38 |  INTELLECT-1 | 34.9% | – | 2024-11-29 |
| 39 |  ChatGLM2-6B | 33.7% | – | 2023-06-24 |
| 40 | LLaMA-7B | 33.5% | – | 2023-02-24 |
| 41 | Baichuan1-7Bopen | 32.5% | – | 2023-06-01 |
| 42 | Falcon-7B | 28.8% | – | 2023-04-24 |
| 43 | Qwen-1_8B | 28.2% | – | 2023-11-30 |
More reasoning benchmarks · All LLM benchmarks
Related benchmarks
Cite as: BenchLeader, “BIG-Bench Hard leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/bbh.
BIG-Bench Hard: questions
- What does BIG-Bench Hard measure?
- Twenty-three hard tasks from BIG-Bench. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on BIG-Bench Hard?
- As of 6 Nov 2026, the highest score on BIG-Bench Hard is 89.2% by Gemini 1.5 Pro 001 (result dated 14 May 2024), ahead of DeepSeek V3 at 87.5%. Scores come from Epoch AI Benchmarking Hub and are re-read daily.
- Which AI model leads BIG-Bench Hard?
- Gemini 1.5 Pro 001 leads BIG-Bench Hard with 89.2% as of 6 Nov 2026 (result dated 14 May 2024), ahead of DeepSeek V3 at 87.5%.
- How many models have BIG-Bench Hard results?
- 43 model configurations have a BIG-Bench Hard result on BenchLeader, all taken from Epoch AI Benchmarking Hub.
- Who runs BIG-Bench Hard and how often is it updated?
- BIG-Bench Hard is published by Epoch AI Benchmarking Hub. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does BIG-Bench Hard count toward the BenchLeader Index?
- No. BIG-Bench Hard is shown for reference but left out of the composite index.