AI benchmarks & best tools
Real, measured benchmark scores for the leading AI models, plus our pick of the best AI tool in each category: coding, image, video, audio, and agents. Refreshed three times a week.
Updated September 11, 2026
LLM benchmark scores
Measured by Artificial Analysis. Higher is better, except price. Sorted by overall intelligence.
| Model | Intelligence | Coding | Math | MMLU-Pro | GPQA | Speed | $/1M |
|---|---|---|---|---|---|---|---|
| Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback)Anthropic | 53 | 82 | 0 | 0% | 94% | 69/s | $20.00 |
| Claude Fable 5.1 (Adaptive Reasoning, Xhigh Effort, Default Fallback)Anthropic | 53 | 81 | 0 | 0% | 93% | 61/s | $20.00 |
| GPT-6 Astra (max)OpenAI | 53 | 77 | 0 | 0% | 96% | 56/s | $20.00 |
| GPT-6 Astra (xhigh)OpenAI | 53 | 76 | 0 | 0% | 96% | 53/s | $20.00 |
| Claude Fable 5.1 (Adaptive Reasoning, High Effort, Default Fallback)Anthropic | 51 | 79 | 0 | 0% | 91% | 53/s | $20.00 |
| GPT-6 Astra (high)OpenAI | 51 | 77 | 0 | 0% | 95% | 53/s | $20.00 |
| Claude Opus 5 (Adaptive Reasoning, Max Effort)Anthropic | 51 | 78 | 0 | 0% | 93% | 58/s | $10.00 |
| GPT-6 Astra (medium)OpenAI | 50 | 77 | 0 | 0% | 94% | 49/s | $20.00 |
| Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)Anthropic | 50 | 77 | 0 | 0% | 93% | 69/s | $20.00 |
| Claude Opus 5 (Adaptive Reasoning, Xhigh Effort)Anthropic | 50 | 77 | 0 | 0% | 94% | 55/s | $10.00 |
| Claude Fable 5.1 (Adaptive Reasoning, Medium Effort, Default Fallback)Anthropic | 49 | 77 | 0 | 0% | 89% | 55/s | $20.00 |
| Muse Spark 1.3 (max)Meta | 48 | 76 | 0 | 0% | 94% | 455/s | $2.00 |
| Claude Opus 5 (Adaptive Reasoning, High Effort)Anthropic | 48 | 77 | 0 | 0% | 94% | 54/s | $10.00 |
| GPT-5.6 Sol (max)OpenAI | 47 | 77 | 0 | 0% | 94% | 72/s | $8.00 |
| Claude Fable 5.1 (Adaptive Reasoning, Low Effort, Default Fallback)Anthropic | 47 | 75 | 0 | 0% | 88% | 51/s | $20.00 |
| GPT-6 Astra (low)OpenAI | 46 | 76 | 0 | 0% | 93% | 49/s | $20.00 |
Benchmark scores: Artificial Analysis · Search interest: DataForSEO · Category analysis: Moonshot AI, curated by Marin AI. Scores reflect public benchmarks and may not capture every real-world strength. Tool picks are our editorial view, refreshed three times a week.
Not sure which tools fit your business?
Knowing the best tool is the easy part. We help you pick the right ones, build them into your workflow, and train your team. Book a free fit call.
Book a free fit call