Task results
Unpublished math problems (old version, FrontierMath 2025-02)
The old version of the set of unpublished math problems, written and vetted by mathematicians, covering advanced undergraduate to early research level.
Superseded by: FrontierMath Tiers 1-3 v2
How to read this table
- Higher rows are rated higher. A range like “1–2” means the gap is within the margin of error.
- The score is the share of a shared set of problems the model solved (%).
- ± is how far the score moves when re-measured, so small gaps do not mean much.
- Prices are a guide per million tokens, reading then writing.
- Length is how much text the model can read at once (tokens).
± range of the top model
| Rank | Model | Developer | Accuracy (%) | Evaluated | Input $ /1M | Output $ /1M | Context | Supports | Weights | Released | Price and specs |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1–5 | GPT-5.5 Pro | OpenAI | Rating 91.9 ±10.0 | $30 | $180 | 1.05M | OpenAI Released 2026-04 Price $30 / $180 Length 1.05M | ||||
| 1–6 | GPT-5.5 | OpenAI | Rating 90.7 ±10.0 | $5 | $30 | 1.05M | OpenAI Released 2026-04 Price $5 / $30 Length 1.05M | ||||
| 1–6 | GPT-5.4 Pro | OpenAI | Rating 87.7 ±10.0 | $30 | $180 | 1.05M | OpenAI Released 2026-03 Price $30 / $180 Length 1.05M | ||||
| 1–8 | GPT-5.4 | OpenAI | Rating 83.5 ±10.0 | $2.5 | $15 | 1.05M | OpenAI Released 2026-03 Price $2.5 / $15 Length 1.05M | ||||
| 1–8 | Claude Opus 4.8 | Anthropic | Rating 82.9 ±10.1 | $5 | $25 | 1M | Anthropic Released 2026-05 Price $5 / $25 Length 1M | ||||
| 2–13 | Claude Opus 4.7 | Anthropic | Rating 76.8 ±10.0 | $5 | $25 | 1M | Anthropic Released 2026-04 Price $5 / $25 Length 1M | ||||
| 4–15 | Claude Opus 4.6 | Anthropic | Rating 71.4 ±10.0 | $5 | $25 | 1M | Anthropic Released 2026-02 Price $5 / $25 Length 1M | ||||
| 4–15 | GPT-5.2 | OpenAI | Rating 71.4 ±10.0 | $1.75 | $14 | 400K | OpenAI Released 2025-12 Price $1.75 / $14 Length 400K | ||||
| 6–17 | Muse Spark | Meta | Rating 68.4 ±10.0 | — | — | — | Meta Released 2026-04 Price — Length — | ||||
| 6–17 | Gemini 3.5 Flash | Rating 68.4 ±9.9 | $1.5 | $9 | 1.05M | Google Released 2026-05 Price $1.5 / $9 Length 1.05M | |||||
| 6–17 | Kimi K2.6 | Moonshot | Rating 68.4 ±9.9 | $0.95 | $4 | 262.1K | Moonshot Released 2026-04 Price $0.95 / $4 Length 262.1K | ||||
| 6–18 | Gemini 3 Pro (preview) | Rating 66.0 ±9.8 | — | — | 1M | Google Released 2025-11 Price — Length 1M | |||||
| 6–18 | Gemini 3.1 Pro (preview) | Rating 64.7 ±9.6 | $2 | $12 | 1.05M | Google Released 2026-02 Price $2 / $12 Length 1.05M | |||||
| 7–19 | Gemini 3 Flash (preview) | Rating 62.5 ±9.7 | $0.5 | $3 | 1M | Google Released 2025-12 Price $0.5 / $3 Length 1M | |||||
| 7–23 | GLM-5.1 | Z.ai | Rating 58.7 ±9.5 | $1.4 | $4.4 | 200K | Z.ai Released 2026-04 Price $1.4 / $4.4 Length 200K | ||||
| 9–24 | GPT-5 | OpenAI | Rating 56.9 ±9.5 | $1.25 | $10 | 400K | OpenAI Released 2025-08 Price $1.25 / $10 Length 400K | ||||
| 9–24 | Claude Sonnet 4.6 | Anthropic | Rating 56.8 ±9.7 | $3 | $15 | 1M | Anthropic Released 2026-02 Price $3 / $15 Length 1M | ||||
| 12–26 | GPT-5.1 | OpenAI | Rating 54.4 ±9.4 | $1.25 | $10 | 400K | OpenAI Released 2025-11 Price $1.25 / $10 Length 400K | ||||
| 14–27 | Gemini 2.5 Deep Think (web) | Rating 50.9 ±9.3 | — | — | — | — | Google Released 2025-08 Price — Length — | ||||
| 15–27 | GPT-5.4 mini | OpenAI | Rating 49.6 ±7.1 | $0.75 | $4.5 | 400K | OpenAI Released 2026-03 Price $0.75 / $4.5 Length 400K | ||||
| 15–29 | Kimi K2.5 | Moonshot | Rating 48.9 ±8.9 | — | — | 262.1K | Moonshot Released 2026-01 Price — Length 262.1K | ||||
| 15–30 | GPT-5 mini | OpenAI | Rating 47.8 ±9.0 | $0.25 | $2 | 400K | OpenAI Released 2025-08 Price $0.25 / $2 Length 400K | ||||
| 15–31 | Qwen3.6 Plus | Alibaba | Rating 46.0 ±8.9 | $0.5 | $3 | 1M | Alibaba Released 2026-03 Price $0.5 / $3 Length 1M | ||||
| 16–31 | GPT-5.4 nano | OpenAI | Rating 45.4 ±8.9 | $0.2 | $1.25 | 400K | OpenAI Released 2026-03 Price $0.2 / $1.25 Length 400K | ||||
| 18–32 | o4-mini | OpenAI | Rating 43.6 ±8.7 | $1.1 | $4.4 | 200K | OpenAI Released 2025-04 Price $1.1 / $4.4 Length 200K | ||||
| 19–32 | Qwen3.6 Max (preview) | Alibaba | Rating 40.5 ±8.5 | $1.3 | $7.8 | 240K | Alibaba Released 2026-04 Price $1.3 / $7.8 Length 240K | ||||
| 18–35 | DeepSeek-V3.2 | DeepSeek | Rating 38.8 ±14.1 | — | — | 128K | DeepSeek Released 2025-12 Price — Length 128K | ||||
| 21–34 | Kimi K2 (thinking) | Moonshot | Rating 37.6 ±8.4 | — | — | 262.1K | Moonshot Released 2025-11 Price — Length 262.1K | ||||
| 21–34 | Qwen3.5 Plus | Alibaba | Rating 36.9 ±8.2 | $0.4 | $2.4 | 1M | Alibaba Released 2026-02 Price $0.4 / $2.4 Length 1M | ||||
| 22–34 | Claude Opus 4.5 | Anthropic | Rating 36.3 ±8.2 | $5 | $25 | 200K | Anthropic Released 2025-11 Price $5 / $25 Length 200K | ||||
| 23–35 | Grok 4 | xAI | Rating 34.5 ±8.0 | — | — | 256K | xAI Released 2025-07 Price — Length 256K | ||||
| 25–35 | o3 | OpenAI | Rating 32.8 ±7.9 | $2 | $8 | 200K | OpenAI Released 2025-04 Price $2 / $8 Length 200K | ||||
| 27–36 | GLM-5 | Z.ai | Rating 28.8 ±7.5 | $1 | $3.2 | 204.8K | Z.ai Released 2026-02 Price $1 / $3.2 Length 204.8K | ||||
| 27–38 | Claude Sonnet 4.5 | Anthropic | Rating 26.7 ±7.3 | $3 | $15 | 200K | Anthropic Released 2025-09 Price $3 / $15 Length 200K | ||||
| 30–39 | Gemini 2.5 Pro | Rating 24.8 ±7.0 | $1.25 | $10 | 1M | Google Released 2025-06 Price $1.25 / $10 Length 1M | |||||
| 33–41 | o3-mini | OpenAI | Rating 21.8 ±6.7 | $1.1 | $4.4 | 200K | OpenAI Released 2025-01 Price $1.1 / $4.4 Length 200K | ||||
| 34–44 | Gemini 2.5 Pro (preview, 2025-05) | Rating 18.1 ±6.2 | — | — | 1.05M | Google Released 2025-05 Price — Length 1.05M | |||||
| 34–44 | Qwen3.6 Flash | Alibaba | Rating 18.1 ±6.2 | $0.25 | $1.5 | 1M | Alibaba Released 2026-04 Price $0.25 / $1.5 Length 1M | ||||
| 35–46 | o1 | OpenAI | Rating 16.3 ±5.9 | $15 | $60 | 128K | OpenAI Released 2024-12 Price $15 / $60 Length 128K | ||||
| 36–49 | Qwen3 235B A22B (thinking) | Alibaba | Rating 14.9 ±5.7 | — | — | 262.1K | Alibaba Released 2025-07 Price — Length 262.1K | ||||
| 36–49 | GPT-5 nano | OpenAI | Rating 14.5 ±5.6 | $0.05 | $0.4 | 400K | OpenAI Released 2025-08 Price $0.05 / $0.4 Length 400K | ||||
| 37–53 | Claude Opus 4.1 | Anthropic | Rating 12.7 ±5.2 | — | — | 200K | Anthropic Released 2025-08 Price — Length 200K | ||||
| 37–53 | Qwen3.5 Flash | Alibaba | Rating 10.9 ±4.9 | $0.1 | $0.4 | 1M | Alibaba Released 2026-02 Price $0.1 / $0.4 Length 1M | ||||
| 37–53 | Claude Haiku 4.5 | Anthropic | Rating 10.4 ±4.8 | $1 | $5 | 200K | Anthropic Released 2025-10 Price $1 / $5 Length 200K | ||||
| 39–53 | Grok 3 Mini (beta) | xAI | Rating 10.3 ±4.8 | — | — | — | xAI Released 2025-04 Price — Length — | ||||
| 39–54 | GPT-4.1 | OpenAI | Rating 9.7 ±4.6 | $2 | $8 | 1.05M | OpenAI Released 2025-04 Price $2 / $8 Length 1.05M | ||||
| 40–55 | Gemini 2.5 Flash | Rating 8.5 ±4.3 | $0.3 | $2.5 | 1M | Google Released 2025-06 Price $0.3 / $2.5 Length 1M | |||||
| 40–59 | Claude Opus 4 | Anthropic | Rating 7.9 ±4.2 | — | — | 200K | Anthropic Released 2025-05 Price — Length 200K | ||||
| 40–59 | GPT-4.1 mini | OpenAI | Rating 7.9 ±4.2 | — | — | 1.05M | OpenAI Released 2025-04 Price — Length 1.05M | ||||
| 42–59 | Claude 3.7 Sonnet | Anthropic | Rating 7.3 ±4.0 | — | — | 200K | Anthropic Released 2025-02 Price — Length 200K | ||||
| 42–59 | Claude Sonnet 4 | Anthropic | Rating 7.3 ±4.0 | — | — | 200K | Anthropic Released 2025-05 Price — Length 200K | ||||
| 42–59 | GLM-4.6 | Z.ai | Rating 6.7 ±3.9 | $0.6 | $2.2 | 204.8K | Z.ai Released 2025-09 Price $0.6 / $2.2 Length 204.8K | ||||
| 42–59 | Grok 3 (beta) | xAI | Rating 6.7 ±3.9 | — | — | — | xAI Released 2025-04 Price — Length — | ||||
| 46–64 | GLM-4.7 | Z.ai | Rating 4.3 ±3.1 | $0.6 | $2.2 | 204.8K | Z.ai Released 2025-12 Price $0.6 / $2.2 Length 204.8K | ||||
| 47–69 | Claude 3.5 Sonnet (2024-10) | Anthropic | Rating 3.6 ±2.9 | — | — | 200K | Anthropic Released 2024-10 Price — Length 200K | ||||
| 48–69 | DeepSeek-V3 | DeepSeek | Rating 3.0 ±2.6 | — | — | 128K | DeepSeek Released 2024-12 Price — Length 128K | ||||
| 48–69 | Gemini 2.0 Flash | Rating 3.0 ±2.6 | — | — | — | Google Released 2025-02 Price — Length — | |||||
| 48–69 | o1-mini | OpenAI | Rating 3.0 ±2.6 | — | — | 128K | OpenAI Released 2024-09 Price — Length 128K | ||||
| 48–69 | Qwen Plus (2025-04) | Alibaba | Rating 3.0 ±2.6 | — | — | — | Alibaba Released 2025-04 Price — Length — | ||||
| 54–71 | Claude 3.5 Sonnet (2024-06) | Anthropic | Rating 1.8 ±2.0 | — | — | 200K | Anthropic Released 2024-06 Price — Length 200K | ||||
| 54–71 | GPT-4.1 nano | OpenAI | Rating 1.8 ±2.0 | $0.1 | $0.4 | 1.05M | OpenAI Released 2025-04 Price $0.1 / $0.4 Length 1.05M | ||||
| 54–71 | Qwen Max | Alibaba | Rating 1.8 ±2.0 | — | — | 128K | Alibaba Released 2025-01 Price — Length 128K | ||||
| 54–71 | Grok 2 (2024-12) | xAI | Rating 1.2 ±1.7 | — | — | — | xAI Released 2024-12 Price — Length — | ||||
| 54–71 | Llama 4 Maverick (FP8) | Meta | Rating 1.2 ±1.7 | — | — | 1.05M | Meta Released 2025-04 Price — Length 1.05M | ||||
| 55–71 | Mistral Medium 3 | Mistral | Rating 0.6 ±1.2 | $0.4 | $2 | 128K | Mistral Released 2025-05 Price $0.4 / $2 Length 128K | ||||
| 55–71 | Claude 3.5 Haiku | Anthropic | Rating 0.6 ±1.2 | — | — | 200K | Anthropic Released 2024-10 Price — Length 200K | ||||
| 55–71 | GPT-4o (2024-08) | OpenAI | Rating 0.6 ±1.2 | $2.5 | $10 | 128K | OpenAI Released 2024-08 Price $2.5 / $10 Length 128K | ||||
| 55–71 | GPT-4o (2024-11) | OpenAI | Rating 0.6 ±1.2 | $2.5 | $10 | 128K | OpenAI Released 2024-11 Price $2.5 / $10 Length 128K | ||||
| 55–71 | Mistral Large (2024-11) | Mistral | Rating 0.6 ±1.2 | — | — | 128K | Mistral Released 2024-11 Price — Length 128K | ||||
| 60–71 | Gemini 1.5 Flash 002 | Rating 0.0 ±0.0 | — | — | — | Google Released 2024-09 Price — Length — | |||||
| 60–71 | Llama 4 Scout | Meta | Rating 0.0 ±0.0 | — | — | 128K | Meta Released 2025-04 Price — Length 128K |
Best accuracy over time
2024-06-20Release date2026-05-28
How to read this page
- Its successor is Tiers 1-3 v2, which fixed problems with errors. The value is accuracy.
- In the Supports column the icons are, from left: can use tools / can read images / can answer in a fixed format (JSON); faded means not supported or unknown.
- Open-weights models have their contents (the weights) published, so you can also run them on your own computer or server. Models without the open-weights mark are normally used through the developer's own services (apps or API); for some models the weights status is unknown.
- Names such as ChatGPT next to a model are the consumer services its developer offers. Check on the service itself whether that model is available there.
Data: Arena, Epoch AI, Swallow LLM team (CC BY 4.0), Models.dev (MIT), updated 2026-10-24, data sources