Skip to main content
Bench

Language

English日本語中文Español

Site info

Task results

Advanced undergraduate to early research math (FrontierMath Tiers 1-3 v2)

The FrontierMath base set of unpublished math problems, written and vetted by mathematicians, covering advanced undergraduate to early research level.

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

RankModelDeveloperAccuracy
(%)Accuracy scale 40 to 100		EvaluatedInput $
/1M		Output $
/1M		ContextSupportsWeightsReleasedPrice and specs
1–6GPT-6.1 SolOpenAIRating 98.6 ±3.0$2$101.05MOpenAI Released 2026-⁠09 Price $2 / $10 Length 1.05M
1–6GPT-6 AstraOpenAIRating 98.6 ±3.0$10$501.05MOpenAI Released 2026-⁠09 Price $10 / $50 Length 1.05M
1–9Claude Opus 5.5AnthropicRating 96.0 ±3.5$4$201MAnthropic Released 2026-⁠09 Price $4 / $20 Length 1M
1–13Claude Fable 5.1AnthropicRating 94.9 ±3.6$10$501MAnthropic Released 2026-⁠09 Price $10 / $50 Length 1M
1–13GPT-6 SolOpenAIRating 94.6 ±3.7$2$101.05MOpenAI Released 2026-⁠09 Price $2 / $10 Length 1.05M
1–13GPT-5.6 SolOpenAIRating 93.8 ±3.8$4$201.05MOpenAI Released 2026-⁠07 Price $4 / $20 Length 1.05M
3–13Claude Sonnet 5.5AnthropicRating 93.4 ±3.9$2$101MAnthropic Released 2026-⁠09 Price $2 / $10 Length 1M
3–16GPT-5.5 ProOpenAIRating 92.3 ±4.0$30$1801.05MOpenAI Released 2026-⁠04 Price $30 / $180 Length 1.05M
3–16Claude Fable 5AnthropicRating 91.6 ±4.1$10$501MAnthropic Released 2026-⁠06 Price $10 / $50 Length 1M
4–17Gemini 4 ArgonGoogleRating 90.5 ±4.3$2$101.05MGoogle Released 2026-⁠09 Price $2 / $10 Length 1.05M
4–17GPT-5.6 TerraOpenAIRating 90.5 ±4.3$2$121.05MOpenAI Released 2026-⁠07 Price $2 / $12 Length 1.05M
4–17Claude Opus 5AnthropicRating 90.1 ±4.3$5$251MAnthropic Released 2026-⁠07 Price $5 / $25 Length 1M
4–17GPT-5.5OpenAIRating 89.8 ±4.3$5$301.05MOpenAI Released 2026-⁠04 Price $5 / $30 Length 1.05M
8–19GPT-5.4 ProOpenAIRating 86.8 ±4.7$30$1801.05MOpenAI Released 2026-⁠03 Price $30 / $180 Length 1.05M
8–19GPT-5.6 LunaOpenAIRating 86.4 ±4.7$0.2$1.21.05MOpenAI Released 2026-⁠07 Price $0.2 / $1.2 Length 1.05M
8–19Muse WonderMetaRating 86.4 ±4.7$4$201.05MMeta Released 2026-⁠10 Price $4 / $20 Length 1.05M
10–23Claude Opus 4.8AnthropicRating 84.2 ±4.9$5$251MAnthropic Released 2026-⁠05 Price $5 / $25 Length 1M
14–24GPT-6 LunaOpenAIRating 83.1 ±5.0$0.1$0.51.05MOpenAI Released 2026-⁠09 Price $0.1 / $0.5 Length 1.05M
14–25GPT-5.4OpenAIRating 82.7 ±5.0$2.5$151.05MOpenAI Released 2026-⁠03 Price $2.5 / $15 Length 1.05M
17–30Claude Haiku 5.5AnthropicRating 79.0 ±5.3$0.1$0.51MAnthropic Released 2026-⁠10 Price $0.1 / $0.5 Length 1M
17–32Qwen3.8 MaxAlibabaRating 78.7 ±5.3$2$61MAlibaba Released 2026-⁠08 Price $2 / $6 Length 1M
17–32Muse Spark 1.3MetaRating 78.3 ±5.3$1.25$4.251.05MMeta Released 2026-⁠09 Price $1.25 / $4.25 Length 1.05M
17–32GPT-5.2 ProOpenAIRating 77.9 ±5.4$21$168400KOpenAI Released 2025-⁠12 Price $21 / $168 Length 400K
18–38Kimi K3⁠MoonshotRating 76.0 ±5.5$3$151.05MMoonshot Released 2026-⁠07 Price $3 / $15 Length 1.05M
19–38Gemini 3.7 FlashGoogleRating 75.3 ±5.5$0.75$3.751.05MGoogle Released 2026-⁠08 Price $0.75 / $3.75 Length 1.05M
20–38DeepSeek-V4.1 Flash (2026-⁠10)⁠DeepSeekRating 75.0 ±5.5$0.15$0.61MDeepSeek Released 2026-⁠10 Price $0.15 / $0.6 Length 1M
20–39Claude Opus 4.7AnthropicRating 73.9 ±5.6$5$251MAnthropic Released 2026-⁠04 Price $5 / $25 Length 1M
20–39Muse Spark Open⁠MetaRating 73.5 ±5.6——1.04MMeta Released 2026-⁠10 Price — Length 1.04M
20–39GLM-5.3⁠Z.aiRating 72.4 ±5.7$1.4$4.41MZ.ai Released 2026-⁠08 Price $1.4 / $4.4 Length 1M
20–39Gemini 3.8 FlashGoogleRating 72.0 ±5.7$0.75$3.751.05MGoogle Released 2026-⁠09 Price $0.75 / $3.75 Length 1.05M
21–40GPT-5.2OpenAIRating 70.9 ±5.8$1.75$14400KOpenAI Released 2025-⁠12 Price $1.75 / $14 Length 400K
21–40DeepSeek-V4.1 Flash⁠DeepSeekRating 70.9 ±5.7——1MDeepSeek Released 2026-⁠09 Price — Length 1M
24–43Claude Opus 4.6AnthropicRating 69.4 ±5.8$5$251MAnthropic Released 2026-⁠02 Price $5 / $25 Length 1M
24–43Grok 4.6xAIRating 69.4 ±5.8$2$6500KxAI Released 2026-⁠08 Price $2 / $6 Length 500K
24–43Claude Sonnet 5AnthropicRating 69.1 ±5.8$2$101MAnthropic Released 2026-⁠06 Price $2 / $10 Length 1M
24–43Qwen3.8 Max (2026-⁠09)AlibabaRating 69.1 ±5.8——1MAlibaba Released 2026-⁠09 Price — Length 1M
24–46DeepSeek-V4 Pro (2026-⁠08)⁠DeepSeekRating 68.0 ±5.9$0.66$1.981MDeepSeek Released 2026-⁠08 Price $0.66 / $1.98 Length 1M
24–46Qwen3.7 MaxAlibabaRating 68.0 ±5.9$2.5$7.51MAlibaba Released 2026-⁠05 Price $2.5 / $7.5 Length 1M
27–49Gemini 3.5 FlashGoogleRating 66.1 ±5.9$1.5$91.05MGoogle Released 2026-⁠05 Price $1.5 / $9 Length 1.05M
31–51Gemini 3.1 Pro (preview)GoogleRating 62.8 ±6.0$2$121.05MGoogle Released 2026-⁠02 Price $2 / $12 Length 1.05M

Best accuracy over time

2024-01-25Release date2026-10-17
How to read this page
  • Models solve them while running Python and submit the answer as a function. 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