Knowledge
SimpleQA Leaderboard
SimpleQA measures factual accuracy on short, fact-seeking questions with a single correct answer — directly probing how often a model is right versus confidently wrong (hallucination) on simple facts.
Source: epoch17 open models ranked+74 proprietaryData through Oct 2026
All models ranked on SimpleQA
Proprietary / closed models are shown dimmed — you can't run them locally, but they show where the open field stands.
| # | Model | Score | VRAM @ Q4 |
|---|---|---|---|
| 1 | Gemini 4 Argon (high) · proprietary | 76.0% | API only |
| 2 | GPT 6 Astra Max · proprietary | 75.6% | API only |
| 3 | GPT 6.1 Sol Max · proprietary | 73.9% | API only |
| 4 | Gemini 3.1 Pro Preview (high) · proprietary | 73.5% | API only |
| 5 | Claude Opus 5.5 Max · proprietary | 72.2% | API only |
| 6 | Claude Fable 5.1 Max · proprietary | 70.8% | API only |
| 7 | Claude Fable 5 (xhigh) · proprietary | 70.7% | API only |
| 8 | Gemini 3.8 Flash (high) · proprietary | 69.7% | API only |
| 9 | GPT 5.6 Sol Max · proprietary | 69.7% | API only |
| 10 | Gemini 3.7 Flash (high) · proprietary | 69.2% | API only |
| 11 | Gemini 3 Flash Preview (high) · proprietary | 66.8% | API only |
| 12 | Gemini 3.5 Flash (high) · proprietary | 66.2% | API only |
| 13 | Gemini 3.6 Flash (high) · proprietary | 66.2% | API only |
| 14 | GPT 5.5 (xhigh) · proprietary | 63.0% | API only |
| 15 | GPT 6 Sol Max · proprietary | 60.7% | API only |
| 16 | Muse Spark 1.2 (xhigh) · proprietary | 60.3% | API only |
| 17 | Claude Opus 5 Max · proprietary | 59.9% | API only |
| 18 | Muse Spark 1.1 · proprietary | 57.8% | API only |
| 19 | Muse Spark Open 816B A42B · 822.0B | 56.6% | ~543 GB |
| 20 | Grok 4.7 (xhigh) · proprietary | 56.0% | API only |
| 21 | Qwen3.7 Max · proprietary | 55.8% | API only |
| 22 | Claude Opus 4.8 Max · proprietary | 53.0% | API only |
| 23 | DeepSeek V4 Pro 0813 · 1650.5B | 52.9% | ~991 GB |
| 24 | Qwen3.6 Max Preview · proprietary | 52.0% | API only |
| 25 | Claude Opus 4.7 (xhigh) · proprietary | 51.7% | API only |
| 26 | Kimi K3 · 2779.9B | 50.6% | ~1674 GB |
| 27 | GPT 5 (Aug 07, 2025, high) · proprietary | 50.1% | API only |
| 28 | O3 (Apr 16, 2025, high) · proprietary | 49.4% | API only |
| 29 | Grok 4.6 (high) · proprietary | 49.3% | API only |
| 30 | Grok 4.6 (xhigh) · proprietary | 48.9% | API only |
| 31 | Qwen3 Max (Sep 23, 2025) · proprietary | 48.8% | API only |
| 32 | Grok 4.5 (high) · proprietary | 48.3% | API only |
| 33 | GPT 5.1 (Nov 13, 2025, high) · proprietary | 48.0% | API only |
| 34 | Qwen3.8 Max (Sep 02, xhigh) · proprietary | 47.3% | API only |
| 35 | Claude Opus 4.6 Max · proprietary | 47.0% | API only |
| 36 | DeepSeek V4 Pro · 1598.8B | 47.0% | ~960 GB |
| 37 | Claude Sonnet 5.5 Max · proprietary | 46.5% | API only |
| 38 | GPT 5.4 Pro (Mar 05, 2026, xhigh) · proprietary | 46.3% | API only |
| 39 | Qwen3.8 Max (xhigh) · proprietary | 45.8% | API only |
| 40 | Claude Opus 4.5 (Nov 01, 2025, 32K) · proprietary | 45.7% | API only |
| 41 | GPT 5.4 (Mar 05, 2026, xhigh) · proprietary | 45.1% | API only |
| 42 | Qwen3.6 Plus · proprietary | 44.1% | API only |
| 43 | GPT 5.6 Terra Max · proprietary | 43.2% | API only |
| 44 | DeepSeek V4.1 Flash · 763.2B | 42.9% | ~458 GB |
| 45 | GPT 6 Luna Max · proprietary | 41.4% | API only |
| 46 | O1 (Dec 17, 2024, high) · proprietary | 41.1% | API only |
| 47 | GLM 5.3 · 753.3B | 41.0% | ~456 GB |
| 48 | GPT 5.6 Luna Max · proprietary | 41.0% | API only |
| 49 | Qwen3 235B A22B Thinking 2507 · 235.1B | 40.4% | ~142 GB |
| 50 | Inkling · 952.4B | 40.3% | ~572 GB |
| 51 | GPT 5.2 (Dec 11, 2025, xhigh) · proprietary | 37.1% | API only |
| 52 | Kimi K2.7 Code · 1026.9B | 36.5% | ~620 GB |
| 53 | Claude Sonnet 4.6 (high) · proprietary | 35.5% | API only |
| 54 | Kimi K2.6 · 1026.9B | 34.9% | ~620 GB |
| 55 | GPT 5.2 (Dec 11, 2025, high) · proprietary | 34.3% | API only |
| 56 | Kimi K2.5 · 1026.9B | 34.3% | ~620 GB |
| 57 | GLM 5.2 · 753.3B | 34.2% | ~456 GB |
| 58 | GLM 5.1 · 753.9B | 34.0% | ~457 GB |
| 59 | Claude Sonnet 5 Max · proprietary | 33.7% | API only |
| 60 | DeepSeek V4 Flash 0731 · 304.2B | 33.6% | ~183 GB |
| 61 | Grok 4.3 (high) · proprietary | 33.2% | API only |
| 62 | Claude Sonnet 5 (xhigh) · proprietary | 32.9% | API only |
| 63 | Claude Sonnet 4.6 Max · proprietary | 32.8% | API only |
| 64 | GPT 5.2 (Dec 11, 2025, low) · proprietary | 32.8% | API only |
| 65 | GPT 5.2 (Dec 11, 2025, medium) · proprietary | 32.7% | API only |
| 66 | GLM 4.7 · 358.3B | 32.2% | ~216 GB |
| 67 | GPT 4.1 (Apr 14, 2025) · proprietary | 31.1% | API only |
| 68 | Claude Sonnet 4.5 (Sep 29, 2025, 59K) · proprietary | 30.7% | API only |
| 69 | Grok 4.20 0309 Reasoning · proprietary | 30.2% | API only |
| 70 | GPT 5.4 Mini (Mar 17, 2026, high) · proprietary | 29.4% | API only |
| 71 | GPT 4o (Aug 06, 2024) · proprietary | 26.0% | API only |
| 72 | Qwen3.5 Plus · proprietary | 25.4% | API only |
| 73 | Claude Haiku 5.5 Max · proprietary | 23.8% | API only |
| 74 | Claude Sonnet 4.5 (Sep 29, 2025) · proprietary | 23.7% | API only |
| 75 | GPT 5 Mini (Aug 07, 2025, high) · proprietary | 21.6% | API only |
| 76 | Qwen3.5 Flash · proprietary | 20.3% | API only |
| 77 | Mistral Large 4.0 (high) · proprietary | 20.0% | API only |
| 78 | O4 Mini (Apr 16, 2025, low) · proprietary | 19.6% | API only |
| 79 | Inkling Small · 266.0B | 19.1% | ~160 GB |
| 80 | O4 Mini (Apr 16, 2025, high) · proprietary | 18.8% | API only |
| 81 | Qwen3.6 Flash · proprietary | 15.9% | API only |
| 82 | O3 Mini (Jan 31, 2025, high) · proprietary | 15.3% | API only |
| 83 | Claude Haiku 4.5 (Oct 01, 2025) · proprietary | 13.2% | API only |
| 84 | GPT 4.1 Mini (Apr 14, 2025) · proprietary | 12.7% | API only |
| 85 | Claude 3 Opus (Feb 29, 2024) · proprietary | 12.6% | API only |
| 86 | Claude Haiku 4.5 (Oct 01, 2025, 32K) · proprietary | 12.6% | API only |
| 87 | GPT 5 Nano (Aug 07, 2025, high) · proprietary | 11.7% | API only |
| 88 | GPT 5.4 Nano (Mar 17, 2026, high) · proprietary | 11.7% | API only |
| 89 | Gemma 4 31B IT · 31.3B | 10.4% | ~20 GB |
| 90 | GPT 4o Mini (Jul 18, 2024) · proprietary | 8.3% | API only |
| 91 | GPT 4.1 Nano (Apr 14, 2025) · proprietary | 6.0% | API only |
Score vs model size
Which models give the most quality for their size — the ones worth running locally.
- Gemma 4 31B IT, 31B, score 10.4% — on the efficiency frontier (best score at its size or smaller).
- Qwen3 235B A22B Thinking 2507, 235B, score 40.4% — on the efficiency frontier (best score at its size or smaller).
- GLM 5.3, 753B, score 41.0% — on the efficiency frontier (best score at its size or smaller).
- DeepSeek V4.1 Flash, 763B, score 42.9% — on the efficiency frontier (best score at its size or smaller).
- Muse Spark Open 816B A42B, 822B, score 56.6% — on the efficiency frontier (best score at its size or smaller).
SimpleQA: frequently asked questions
- What is the best open LLM on SimpleQA?
- Muse Spark Open 816B A42B is the top open model on SimpleQA, scoring 56.6%. Among all models tested — including proprietary ones — it ranks #19. The top model overall is Gemini 4 Argon (high) (Google DeepMind) at 76.0%.
- What's the best SimpleQA model you can run on a 24 GB GPU?
- Gemma 4 31B IT is the highest-scoring open model that fits in 24 GB at 4-bit quantization (~20 GB), scoring 10.4% on SimpleQA.
- Can open models match proprietary models on SimpleQA?
- Not quite on SimpleQA: the strongest proprietary model (Gemini 4 Argon (high)) scores 76.0%, ahead of the best open model (Muse Spark Open 816B A42B) at 56.6% — but you can run the open one yourself.
Scores aggregated from epoch. llmrun does not run this benchmark — see the source for methodology, or the about benchmarks for what it measures.