Reported by others
Reading long stories (Fiction.LiveBench)
The model reads a long story and answers questions about how the characters relate and how events unfold.
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 value published by the task's authors and other organizations.
- Conditions can differ between models, 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).
| Rank | Model | Developer | Accuracy (%) | Input $ /1M | Output $ /1M | Context | Supports | Weights | Released | Price and specs |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | o3 | OpenAI | Rating 100.0 | $2 | $8 | 200K | OpenAI Released 2025-04 Price $2 / $8 Length 200K | |||
| 2–3 | GPT-5 | OpenAI | Rating 96.9 | $1.25 | $10 | 400K | OpenAI Released 2025-08 Price $1.25 / $10 Length 400K | |||
| 2–3 | Grok 4 | xAI | Rating 96.9 | — | — | 256K | xAI Released 2025-07 Price — Length 256K | |||
| 4 | Gemini 2.5 Pro (exp) | Rating 90.6 | — | — | 1.05M | Google Released 2025-03 Price — Length 1.05M | ||||
| 5 | o3-pro | OpenAI | Rating 88.9 | — | — | 200K | OpenAI Released 2025-06 Price — Length 200K | |||
| 6 | Gemini 2.5 Pro (preview, 2025-05) | Rating 87.5 | — | — | 1.05M | Google Released 2025-05 Price — Length 1.05M | ||||
| 7 | Kimi K2.5 | Moonshot | Rating 78.1 | — | — | 262.1K | Moonshot Released 2026-01 Price — Length 262.1K | |||
| 8 | Grok 4 Fast | xAI | Rating 75.0 | — | — | 2M | xAI Released 2025-09 Price — Length 2M | |||
| 9 | Gemini 2.5 Pro (preview, 2025-03) | Rating 71.9 | — | — | 1.05M | Google Released 2025-03 Price — Length 1.05M | ||||
| 10–11 | Qwen3 235B A22B (thinking) | Alibaba | Rating 68.8 | — | — | 262.1K | Alibaba Released 2025-07 Price — Length 262.1K | |||
| 10–11 | Gemini 2.5 Flash (preview, 2025-05) | Rating 68.8 | — | — | 1.05M | Google Released 2025-05 Price — Length 1.05M | ||||
| 12–13 | ChatGPT-4o (2025-01) | OpenAI | Rating 65.6 | — | — | — | OpenAI Released 2025-01 Price — Length — | |||
| 12–13 | Grok 3 Mini (beta) | xAI | Rating 65.6 | — | — | — | xAI Released 2025-04 Price — Length — | |||
| 14 | GPT-4.5 (preview) | OpenAI | Rating 63.9 | — | — | — | OpenAI Released 2025-02 Price — Length — | |||
| 15–19 | Qwen3 Next 80B A3B (instruct) | Alibaba | Rating 62.5 | $0.5 | $2 | 129K | Alibaba Released 2025-09 Price $0.5 / $2 Length 129K | |||
| 15–19 | Gemini 2.0 Flash | Rating 62.5 | — | — | — | Google Released 2025-02 Price — Length — | ||||
| 15–19 | GPT-4.1 | OpenAI | Rating 62.5 | $2 | $8 | 1.05M | OpenAI Released 2025-04 Price $2 / $8 Length 1.05M | |||
| 15–19 | GPT-5 mini | OpenAI | Rating 62.5 | $0.25 | $2 | 400K | OpenAI Released 2025-08 Price $0.25 / $2 Length 400K | |||
| 15–19 | o4-mini | OpenAI | Rating 62.5 | $1.1 | $4.4 | 200K | OpenAI Released 2025-04 Price $1.1 / $4.4 Length 200K | |||
| 20 | MiniMax-M1-80k | MiniMax | Rating 59.4 | — | — | 1M | MiniMax Released 2025-06 Price — Length 1M | |||
| 21 | Grok 3 (beta) | xAI | Rating 58.3 | — | — | — | xAI Released 2025-04 Price — Length — | |||
| 22–25 | Claude 3.7 Sonnet | Anthropic | Rating 53.1 | — | — | 200K | Anthropic Released 2025-02 Price — Length 200K | |||
| 22–25 | DeepSeek-V3.1 | DeepSeek | Rating 53.1 | — | — | 128K | DeepSeek Released 2025-08 Price — Length 128K | |||
| 22–25 | Gemini 2.5 Flash (preview, 2025-04) | Rating 53.1 | — | — | 1.05M | Google Released 2025-04 Price — Length 1.05M | ||||
| 22–25 | o1 | OpenAI | Rating 53.1 | $15 | $60 | 128K | OpenAI Released 2024-12 Price $15 / $60 Length 128K | |||
| 26 | GPT-4.1 mini | OpenAI | Rating 46.9 | — | — | 1.05M | OpenAI Released 2025-04 Price — Length 1.05M | |||
| 27 | Qwen3 235B A22B (instruct) | Alibaba | Rating 44.4 | — | — | 128K | Alibaba Released 2025-07 Price — Length 128K | |||
| 28 | o3-mini | OpenAI | Rating 43.8 | $1.1 | $4.4 | 200K | OpenAI Released 2025-01 Price $1.1 / $4.4 Length 200K | |||
| 29 | Kimi K2 (instruct, 2025-09) | Moonshot | Rating 40.6 | — | — | 262.1K | Moonshot Released 2025-09 Price — Length 262.1K | |||
| 30–32 | Claude Opus 4 | Anthropic | Rating 37.5 | — | — | 200K | Anthropic Released 2025-05 Price — Length 200K | |||
| 30–32 | Gemini 2.0 Flash (thinking, exp) | Rating 37.5 | — | — | — | Google Released 2024-12 Price — Length — | ||||
| 30–32 | Gemini 2.0 Pro (exp) | Rating 37.5 | — | — | 2.1M | Google Released 2025-02 Price — Length 2.1M | ||||
| 33–34 | Llama 4 Maverick | Meta | Rating 36.4 | — | — | 524.3K | Meta Released 2025-04 Price — Length 524.3K | |||
| 33–34 | Claude Sonnet 4 | Anthropic | Rating 36.4 | — | — | 200K | Anthropic Released 2025-05 Price — Length 200K | |||
| 35 | DeepSeek-R1 | DeepSeek | Rating 33.3 | — | — | 128K | DeepSeek Released 2025-01 Price — Length 128K | |||
| 36 | Llama 4 Scout | Meta | Rating 27.3 | — | — | 128K | Meta Released 2025-04 Price — Length 128K | |||
| 37–38 | Gemini 2.5 Flash-Lite (preview, 2025-06) | Rating 21.9 | — | — | 1.05M | Google Released 2025-06 Price — Length 1.05M | ||||
| 37–38 | GPT-5 nano | OpenAI | Rating 21.9 | $0.05 | $0.4 | 400K | OpenAI Released 2025-08 Price $0.05 / $0.4 Length 400K | |||
| 39 | GPT-4.1 nano | OpenAI | Rating 18.8 | $0.1 | $0.4 | 1.05M | OpenAI Released 2025-04 Price $0.1 / $0.4 Length 1.05M |
Best accuracy over time
2024-12-17Release date2026-01-27
How to read this page
- The value on this page is accuracy at a length of about 120,000 tokens. Figures published by the organizations that made the tasks, among others, collected here.
- The accuracy is the share of problems solved. A gap of 1 point is 1 question out of 100.
- 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