FrontierMath Tiers 1-3 v2 (Epoch AI)
91 models with a score. Leader: GPT-6 astra with 93.7 %.
| Benchmark profile | |
|---|---|
| Task | allgemein - Mathematik bis Forschungsniveau |
| Metric | Anteil geloester Aufgaben, bester Wert ueber die Bewerter |
| Direction | ↑ higher is better |
| Limitation | Fuer die Anreicherung das fernste Signal. Stark reasoning-abhaengig; GPT-5.4 und DeepSeek V4 Pro sind nur mit xhigh/max gemessen. Aufgaben nicht oeffentlich - nicht nachpruefbar, aber auch nicht kontaminiert. |
| As of | 15 Sept 2026 |
| Source | epoch.ai |
| Licence | CC BY 4.0 — Epoch AI |
Ranking (91)
No entry matches the filter.
| Model | Rank | Score | Developer | EU processing: InputEU in | EU processing: OutputEU out | all providers: Inputall in | all providers: Outputall out | Open weights | GPU memory (4-bit, derived)GPU memory |
|---|---|---|---|---|---|---|---|---|---|
| GPT-6 astra | 1 | 93.7 % | OpenAI | €9.68 | €48.40 | €4.48 | €22.39 | – no | – |
| GPT-6.1 Sol | 1 | 93.7 % | OpenAI | €2.14 | €10.68 | €0.90 | €4.48 | – no | – |
| Claude Opus 5.5 | 3 | 91.2 % | Anthropic | €3.92 | €19.58 | €1.79 | €8.96 | – no | – |
| GPT-6.1 Terra | 4 | 90.9 % | OpenAI | €1.06 | €5.28 | €0.45 | €2.24 | – | – |
| Claude Fable 5.1 | 5 | 90.2 % | Anthropic | €9.90 | €49.50 | €4.48 | €22.39 | – no | – |
| GPT-6 sol | 6 | 89.8 % | OpenAI | €2.15 | €10.75 | €0.90 | €4.48 | – no | – |
| GPT-5.6 sol | 7 | 89.1 % | OpenAI | €3.94 | €19.70 | €1.79 | €8.96 | – no | – |
| Sonnet 5.5 | 8 | 88.8 % | Anthropic | €2.02 | €10.10 | €0.90 | €4.48 | – no | – |
| GPT-5.5 Pro | 9 | 87.7 % | OpenAI | – | – | €13.43 | €80.61 | – no | – |
| Fable 5 | 10 | 87 % | Anthropic | – | – | €4.48 | €22.39 | – no | – |
| Gemini 4 Argon | 11 | 86 % | €2.02 | €10.10 | €0.90 | €4.48 | – no | – | |
| GPT-5.6 terra | 11 | 86 % | OpenAI | €1.97 | €11.82 | €0.90 | €5.37 | – no | – |
| Opus 5 | 13 | 85.6 % | Anthropic | €4.94 | €24.68 | €2.24 | €11.20 | – no | – |
| GPT-5.5 | 14 | 85.3 % | OpenAI | €4.93 | €29.56 | €2.24 | €13.43 | – no | – |
| GPT-5.4 Pro | 15 | 82.5 % | OpenAI | – | – | €13.43 | €80.61 | – no | – |
| GPT-5.6 luna | 16 | 82.1 % | OpenAI | €0.20 | €1.18 | €0.09 | €0.54 | – no | – |
| Muse Wonder | 16 | 82.1 % | – | – | – | €3.58 | €17.91 | – no | – |
| Opus 4.8 | 18 | 80 % | Anthropic | €4.88 | €24.40 | €2.24 | €11.20 | – no | – |
| GPT-6 luna | 19 | 78.9 % | OpenAI | €0.11 | €0.54 | €0.045 | €0.22 | – no | – |
| GPT-5.4 | 20 | 78.6 % | OpenAI | €2.60 | €13.87 | €1.12 | €6.72 | – no | – |
| Kimi K3.1 Pro | 21 | 78.2 % | Moonshot AI | €1.79 | €8.96 | €1.79 | €8.96 | – no | – |
| Haiku 5.5 | 22 | 75.1 % | Anthropic | €0.10 | €0.51 | €0.045 | €0.22 | – no | – |
| Qwen3.8-Max | 23 | 74.7 % | Alibaba Cloud | €2.24 | €5.39 | €1.48 | €4.43 | ✓ yes | from 1,350 GB |
| gpt-5.2-pro | 24 | 74 % | OpenAI | – | – | €9.40 | €75.24 | – no | – |
| Kimi K3 | 25 | 72.2 % | Moonshot AI | €2.55 | €12.76 | €1.66 | €8.51 | ✓ yes | from 1,533 GB |
| Gemini 3.7 Flash | 26 | 71.6 % | €0.67 | €3.37 | €0.34 | €1.68 | – no | – | |
| DeepSeek V4.1 Flash 1017 | 27 | 71.2 % | DeepSeek | €0.18 | €0.54 | €0.18 | €0.54 | – no | – |
| Opus 4.7 | 28 | 70.2 % | Anthropic | €4.88 | €24.40 | €2.24 | €11.20 | – no | – |
| Muse Spark Open | 29 | 69.8 % | – | – | – | €1.12 | €3.81 | – | – |
| glm-5.3 | 30 | 68.8 % | Z.ai | €0.45 | €1.79 | €0.18 | €0.57 | ✓ yes | from 419 GB |
| Gemini 3.8 Flash | 31 | 68.4 % | €0.74 | €3.68 | €0.34 | €1.68 | – no | – | |
| GPT-5.2 | 32 | 67.4 % | OpenAI | €1.72 | €13.79 | €0.78 | €6.27 | – no | – |
| Grok 4.6 | 33 | 66 % | xAI | – | – | €1.79 | €5.37 | – no | – |
| Opus 4.6 | 33 | 66 % | Anthropic | €4.77 | €23.84 | €2.24 | €11.20 | – no | – |
| Claude Sonnet 5 | 35 | 65.6 % | Anthropic | €1.98 | €9.87 | €0.90 | €4.48 | – no | – |
| DeepSeek V4 Pro 0813 | 36 | 64.6 % | DeepSeek | €1.79 | €3.58 | €1.16 | €2.33 | ✓ yes | from 912 GB |
| Qwen 3.7 Max | 36 | 64.6 % | Alibaba Cloud | – | – | €1.12 | €3.36 | – no | – |
| Gemini 3.5 Flash | 38 | 62.8 % | €1.48 | €8.89 | €0.67 | €4.03 | – no | – | |
| Gemini 3.1 Pro (Vorschau) | 39 | 59.6 % | – | – | €0.90 | €5.37 | – no | – | |
| glm-5.2 | 40 | 59.2 % | Z.ai | €0.81 | €2.91 | €0.15 | €0.48 | ✓ yes | from 419 GB |
| Gemini 3.6 Flash | 41 | 58.9 % | €0.67 | €3.37 | €0.34 | €1.68 | – no | – | |
| DeepSeek V4 Flash 0731 | 42 | 57.5 % | DeepSeek | €0.10 | €0.25 | €0.054 | €0.16 | ✓ yes | from 172 GB |
| Grok 4.5 | 43 | 57.2 % | xAI | – | – | €1.79 | €5.37 | – no | – |
| Kimi K2.6 | 43 | 57.2 % | Moonshot AI | €0.43 | €2.67 | €0.35 | €1.61 | ✓ yes | from 569 GB |
| GLM 5.3 Flash | 45 | 55.8 % | Z.ai | €0.09 | €0.31 | €0.09 | €0.31 | ✓ yes | from 180 GB |
| GPT-5 Pro | 45 | 55.8 % | OpenAI | – | – | €6.72 | €53.74 | – no | – |
| gpt-5 | 47 | 55.4 % | OpenAI | €0.62 | €4.93 | €0.56 | €4.48 | – no | – |
| GPT-5 2025-08-07 | 47 | 55.4 % | OpenAI | – | – | €0.56 | €4.48 | – no | – |
| Kimi K2.7 Code | 49 | 54 % | Moonshot AI | €0.59 | €2.96 | €0.59 | €2.96 | ✓ yes | from 554 GB |
| Grok 4.7 | 50 | 53 % | xAI | – | – | €1.79 | €5.37 | – no | – |
| Gemini 3 Flash Preview | 51 | 51.2 % | – | – | €0.22 | €1.34 | – no | – | |
| GPT-5 mini | 52 | 46.7 % | OpenAI | €0.25 | €1.97 | €0.11 | €0.90 | – no | – |
| DeepSeek V4 Pro | 53 | 45.3 % | DeepSeek | €0.90 | €1.79 | €0.90 | €1.79 | ✓ yes | from 884 GB |
| Grok 4.20 | 54 | 44.9 % | xAI | – | – | €1.27 | €2.53 | – no | – |
| Grok 4.20 Reasoning | 54 | 44.9 % | xAI | – | – | €1.12 | €2.24 | – no | – |
| grok-4.20-0309-reasoning | 54 | 44.9 % | xAI | – | – | €1.12 | €2.24 | – no | – |
| Grok 4.3 | 57 | 42.8 % | xAI | – | – | €1.12 | €2.24 | – no | – |
| Qwen 3.6 Plus | 58 | 38.2 % | Alibaba Cloud | – | – | €0.45 | €2.69 | – no | – |
| GLM 5.1 | 59 | 36.8 % | Z.ai | €1.24 | €3.90 | €0.81 | €2.60 | ✓ yes | from 419 GB |
| o4-mini 2025-04-16 | 60 | 36.1 % | OpenAI | €0.54 | €2.17 | €0.49 | €1.97 | – no | – |
| Qwen3.6-27B | 61 | 35.09 % | Alibaba Cloud | €0.40 | €2.70 | €0.22 | €1.79 | ✓ yes | from 20 GB |
| Opus 4.5 | 62 | 34.4 % | Anthropic | €4.77 | €23.85 | €2.24 | €11.20 | – no | – |
| Qwen3.7-Plus | 63 | 34.39 % | Alibaba Cloud | – | – | €0.29 | €1.15 | – no | – |
| o3 | 64 | 33.3 % | OpenAI | – | – | €0.90 | €3.58 | – no | – |
| o3 2025-04-16 | 64 | 33.3 % | OpenAI | €0.99 | €3.94 | €0.90 | €3.58 | – no | – |
| Qwen3.5-397B-A17B | 66 | 31.2 % | Alibaba Cloud | €0.60 | €3.58 | €0.40 | €2.69 | ✓ yes | from 226 GB |
| Gemini 3.5 Flash-Lite | 67 | 26 % | €0.30 | €2.47 | €0.13 | €1.12 | – no | – | |
| Gemini 2.5 Pro | 68 | 24.6 % | €1.34 | €8.94 | €0.56 | €4.48 | – no | – | |
| GPT-5.4 mini | 69 | 24.56 % | OpenAI | €0.74 | €4.43 | €0.34 | €2.02 | – no | – |
| Sonnet 4.5 | 70 | 23.9 % | Anthropic | €2.68 | €13.42 | €1.34 | €6.72 | – no | – |
| Gemini 3.1 Flash-Lite | 71 | 21.4 % | €0.24 | €1.46 | €0.11 | €0.67 | – no | – | |
| Qwen3.6-35B-A3B | 72 | 20.4 % | Alibaba Cloud | €0.13 | €0.45 | €0.13 | €0.45 | ✓ yes | from 24 GB |
| GPT-5.4 nano | 73 | 20.35 % | OpenAI | – | – | €0.09 | €0.56 | – no | – |
| GPT-5 nano | 74 | 20 % | OpenAI | €0.049 | €0.39 | €0.022 | €0.18 | – no | – |
| Qwen3-Max | 75 | 18.95 % | Alibaba Cloud | – | – | – | – | – no | – |
| o3 mini 2025-01-31 | 76 | 18.6 % | OpenAI | €0.54 | €2.17 | €0.49 | €1.97 | – no | – |
| Qwen3.5-Flash | 77 | 18.25 % | Alibaba Cloud | – | – | – | – | – no | – |
| GLM-5 | 78 | 16.4 % | Z.ai | €1.15 | €3.68 | €0.45 | €1.43 | – | – |
| o1 2024-12-17 | 79 | 14.7 % | OpenAI | €14.78 | €59.11 | €13.43 | €53.74 | – no | – |
| Claude Opus 4.1 (retired, except on Claude Models on Google Cloud (Google managed) and Bedrock) | 80 | 12.6 % | Anthropic | – | – | €6.72 | €33.59 | – no | – |
| gpt-4.1-mini | 81 | 6.7 % | OpenAI | €0.39 | €1.53 | €0.18 | €0.72 | – no | – |
| GPT-4.1-mini-2025-04-14 | 81 | 6.7 % | OpenAI | €0.20 | €0.79 | €0.18 | €0.72 | – no | – |
| GPT-4.1 | 83 | 6 % | OpenAI | €1.97 | €7.87 | €0.90 | €3.58 | – no | – |
| GPT-4.1-2025-04-14 | 83 | 6 % | OpenAI | €0.99 | €3.94 | €0.90 | €3.58 | – no | – |
| Claude Haiku 4.5 | 85 | 5.9 % | Anthropic | €0.89 | €4.47 | €0.45 | €2.24 | – no | – |
| GPT-4o mini | 86 | 0.7 % | OpenAI | €0.14 | €0.57 | €0.067 | €0.27 | – no | – |
| GPT-4o-mini-0718 | 86 | 0.7 % | OpenAI | €0.15 | €0.59 | €0.067 | €0.27 | – no | – |
| GPT-4o | 88 | 0.4 % | OpenAI | €2.39 | €9.55 | €1.12 | €4.48 | – no | – |
| GPT-4o-2024-08-06 | 88 | 0.4 % | OpenAI | €1.23 | €4.93 | €1.12 | €4.48 | – no | – |
| Mistral Large 2.1 | 90 | 0.34 % | Mistral AI | – | – | – | – | ✓ yes | from 72 GB |
| GPT-3.5-Turbo-0125 | 91 | 0 % | OpenAI | – | – | €0.54 | €1.63 | – no | – |
Equal scores share a rank. When filtering, the rank stays the position in the whole list. GPU memory “from …” is a lower bound (parameters × 0.55 GB + 4 GB). Prices: € per 1M tokens. Source: providers' pricing pages, read and checked by searchit. As of 7 Nov 2026.
Price and performance
Scores: FrontierMath Tiers 1-3 v2 (Epoch AI), CC BY 4.0 — Epoch AI. Taken unchanged from the source.