Skip to main content
Bench

Language

English日本語中文Español

Site info

Task results

Expert-level biology, chemistry, physics questions (the subset only experts answered correctly, GPQA diamond)

Multiple-choice questions written by experts in biology, chemistry, and physics, limited to those that domain experts answered correctly and most non-experts got wrong.

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 85 to 100		EvaluatedInput $
/1M		Output $
/1M		ContextSupportsWeightsReleasedPrice and specs
1–25Gemini 4 ArgonGoogleRating 99.2 ±3.2$2$101.05MGoogle Released 2026-⁠09 Price $2 / $10 Length 1.05M
1–28GPT-6 AstraOpenAIRating 98.3 ±3.7$10$501.05MOpenAI Released 2026-⁠09 Price $10 / $50 Length 1.05M
1–33Claude Sonnet 5.5AnthropicRating 98.0 ±3.7$2$101MAnthropic Released 2026-⁠09 Price $2 / $10 Length 1M
1–34Gemini 3.8 FlashGoogleRating 97.8 ±3.8$0.75$3.751.05MGoogle Released 2026-⁠09 Price $0.75 / $3.75 Length 1.05M
1–34GPT-6.1 SolOpenAIRating 97.8 ±3.8$2$101.05MOpenAI Released 2026-⁠09 Price $2 / $10 Length 1.05M
1–41Gemini 3.7 FlashGoogleRating 97.0 ±3.7$0.75$3.751.05MGoogle Released 2026-⁠08 Price $0.75 / $3.75 Length 1.05M
1–46GPT-5.4 ProOpenAIRating 96.7 ±4.4$30$1801.05MOpenAI Released 2026-⁠03 Price $30 / $180 Length 1.05M
1–47Gemini 3.1 Pro (preview)GoogleRating 96.4 ±4.4$2$121.05MGoogle Released 2026-⁠02 Price $2 / $12 Length 1.05M
1–46GPT-6 SolOpenAIRating 96.3 ±4.1$2$101.05MOpenAI Released 2026-⁠09 Price $2 / $10 Length 1.05M
1–47Gemini 3.6 FlashGoogleRating 96.0 ±3.8$0.75$3.751.05MGoogle Released 2026-⁠07 Price $0.75 / $3.75 Length 1.05M
1–48GPT-5.5OpenAIRating 95.8 ±4.2$5$301.05MOpenAI Released 2026-⁠04 Price $5 / $30 Length 1.05M
1–47Grok 4.6xAIRating 95.8 ±3.9$2$6500KxAI Released 2026-⁠08 Price $2 / $6 Length 500K
1–48GPT-5.5 ProOpenAIRating 95.7 ±4.3$30$1801.05MOpenAI Released 2026-⁠04 Price $30 / $180 Length 1.05M
1–48Claude Opus 5AnthropicRating 95.7 ±4.0$5$251MAnthropic Released 2026-⁠07 Price $5 / $25 Length 1M
1–50GPT-5.6 SolOpenAIRating 95.1 ±4.3$4$201.05MOpenAI Released 2026-⁠07 Price $4 / $20 Length 1.05M
1–50Grok 4.5xAIRating 95.0 ±3.9$2$6500KxAI Released 2026-⁠07 Price $2 / $6 Length 500K
1–54GPT-5.6 TerraOpenAIRating 94.9 ±4.2$2$121.05MOpenAI Released 2026-⁠07 Price $2 / $12 Length 1.05M
1–56GPT-5.4OpenAIRating 94.9 ±4.9$2.5$151.05MOpenAI Released 2026-⁠03 Price $2.5 / $15 Length 1.05M
1–55DeepSeek-V4.1 Flash (2026-⁠10)⁠DeepSeekRating 94.7 ±4.3$0.15$0.61MDeepSeek Released 2026-⁠10 Price $0.15 / $0.6 Length 1M
1–55Kimi K3⁠MoonshotRating 94.6 ±4.1$3$151.05MMoonshot Released 2026-⁠07 Price $3 / $15 Length 1.05M
1–58Gemini 3.5 FlashGoogleRating 94.2 ±4.5$1.5$91.05MGoogle Released 2026-⁠05 Price $1.5 / $9 Length 1.05M
1–58Grok 4.7xAIRating 94.1 ±4.2$2$6500KxAI Released 2026-⁠09 Price $2 / $6 Length 500K
1–58Qwen3.8 MaxAlibabaRating 94.0 ±4.6$2$61MAlibaba Released 2026-⁠08 Price $2 / $6 Length 1M
1–58Gemini 3 Pro (preview)GoogleRating 93.9 ±4.5——1MGoogle Released 2025-⁠11 Price — Length 1M
1–62Qwen3.8 Max (2026-⁠09)AlibabaRating 93.5 ±4.7——1MAlibaba Released 2026-⁠09 Price — Length 1M
2–64GLM-5.2⁠Z.aiRating 92.9 ±4.4$1.4$4.41MZ.ai Released 2026-⁠06 Price $1.4 / $4.4 Length 1M
3–64DeepSeek-V4 Pro (2026-⁠08)⁠DeepSeekRating 92.6 ±4.1$0.66$1.981MDeepSeek Released 2026-⁠08 Price $0.66 / $1.98 Length 1M
2–66GPT-5.6 LunaOpenAIRating 92.5 ±4.7$0.2$1.21.05MOpenAI Released 2026-⁠07 Price $0.2 / $1.2 Length 1.05M
2–69GPT-5.2OpenAIRating 92.2 ±4.9$1.75$14400KOpenAI Released 2025-⁠12 Price $1.75 / $14 Length 400K
3–72Claude Opus 4.8AnthropicRating 91.7 ±5.2$5$251MAnthropic Released 2026-⁠05 Price $5 / $25 Length 1M
4–69DeepSeek-V4 Flash (2026-⁠07)⁠DeepSeekRating 91.7 ±4.8——1MDeepSeek Released 2026-⁠07 Price — Length 1M
3–73DeepSeek-V4 Pro⁠DeepSeekRating 91.5 ±5.6$0.66$1.981MDeepSeek Released 2026-⁠04 Price $0.66 / $1.98 Length 1M
6–69GLM-5.3⁠Z.aiRating 91.5 ±4.3$1.4$4.41MZ.ai Released 2026-⁠08 Price $1.4 / $4.4 Length 1M
3–73MiniMax-M3⁠MiniMaxRating 91.5 ±5.6$0.3$1.21.05MMiniMax Released 2026-⁠06 Price $0.3 / $1.2 Length 1.05M
3–73Qwen3.7 MaxAlibabaRating 91.5 ±5.6$2.5$7.51MAlibaba Released 2026-⁠05 Price $2.5 / $7.5 Length 1M
6–72Kimi K2.6⁠MoonshotRating 91.4 ±4.7$0.95$4262.1KMoonshot Released 2026-⁠04 Price $0.95 / $4 Length 262.1K
6–73Claude Opus 5.5AnthropicRating 91.1 ±5.3$4$201MAnthropic Released 2026-⁠09 Price $4 / $20 Length 1M
7–73Claude Opus 4.6AnthropicRating 91.0 ±4.6$5$251MAnthropic Released 2026-⁠02 Price $5 / $25 Length 1M
6–73Claude Sonnet 5AnthropicRating 91.0 ±4.9$2$101MAnthropic Released 2026-⁠06 Price $2 / $10 Length 1M
6–73GPT-6 LunaOpenAIRating 90.9 ±5.0$0.1$0.51.05MOpenAI Released 2026-⁠09 Price $0.1 / $0.5 Length 1.05M

Best accuracy over time

2023-03-14Release date2026-10-17
How to read this page
  • 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