Best AI Models for Science
Ranking of language models for scientific work — physics, chemistry, biology and advanced mathematics problem solving.Updated August 2026. Based on independent benchmark data.
Why This Matters
Choosing the right model for science is about more than raw benchmark scores. The models below are ranked using a weighted blend of category-specific benchmarks, arena Elo, and the intelligence index — but real-world performance depends on your context window, latency budget, and integration path.
For science, we weighted the benchmarks that most strongly predict success in production: task accuracy first, then speed and cost efficiency. Leaders on this page consistently outperform on both standardized tests and real-world workloads.
If you operate under a strict budget, check the Best Value LLMs ranking. If latency matters more than raw quality, see Fastest LLMs. Both pages use the same underlying dataset with different weightings.
| # | Model | Provider | Score | Speed | Price (output) |
|---|---|---|---|---|---|
| 1 | Claude Mythos Preview | Anthropic | 99.4 | 80 t/s | $15/M |
| 2 | Gemini 3.1 Pro | 85.6 | 136 t/s | $12/M | |
| 3 | Claude Fable 5.1 | Anthropic | 75.0 | 85 t/s | $15/M |
| 4 | Claude Mythos 5 | Anthropic | 73.0 | 50 t/s | $50/M |
| 5 | GPT-6 Astra (xhigh) | OpenAI | 67.5 | 76 t/s | $50/M |
| 6 | GPT-6 Astra (high) | OpenAI | 67.5 | 90 t/s | $50/M |
| 7 | GPT-6 Astra (medium) | OpenAI | 67.5 | 98 t/s | $50/M |
| 8 | GPT-6 Astra (low) | OpenAI | 67.5 | 110 t/s | $50/M |
| 9 | Claude Opus 5.5 | Anthropic | 65.5 | 68 t/s | $25/M |
| 10 | Qwen3.7 Max | Alibaba | 65.2 | 203 t/s | $7.5/M |
| 11 | Claude Fable 5.1 (XHigh) | Anthropic | 65.0 | 78 t/s | $15/M |
| 12 | Claude Fable 5.1 (High) | Anthropic | 65.0 | 92 t/s | $15/M |
| 13 | Claude Fable 5.1 (Medium) | Anthropic | 65.0 | 104 t/s | $15/M |
| 14 | Claude Fable 5.1 (Low) | Anthropic | 65.0 | 118 t/s | $15/M |
| 15 | Qwen3.8 Flash Next | Alibaba | 65.0 | 168 t/s | $0.4/M |
| 16 | Gemini 3 Pro | 64.6 | 120 t/s | $5/M | |
| 17 | Claude Sonnet 4.6 (max) | Anthropic | 64.5 | 50 t/s | $15/M |
| 18 | GPT-6 Astra | OpenAI | 63.8 | 82 t/s | $50/M |
| 19 | GPT-6 Luna (xhigh) | OpenAI | 62.5 | 118 t/s | $20/M |
| 20 | GPT-6 Luna (high) | OpenAI | 62.5 | 128 t/s | $20/M |
Top 5 Science — Quick Comparison
Methodology
Rankings are based on independent benchmark evaluations. The science score is computed from multiple standardized benchmarks and normalized to 0–100. Data is updated regularly from public benchmark results.
Related Rankings
Frequently Asked Questions
Which AI model is best for scientific research?
Top reasoning models scoring highest on GPQA Diamond and graduate-level physics/chemistry/biology benchmarks lead this ranking.