Read
Choosing an LLM for classification and extraction
There is no public evaluation of classification or extraction yet, so this is ranked by human ratings on instruction following.
- Highest rated
- △Claude Opus 5.5 input $4
- On a budget
- GLM-5.3 Flash Z.ai Open weights input $0.15
- Rating gap 48
- When people compare the two answers, Claude Opus 5.5 is picked about 57% of the time
About the same rating as the cheaper Gemini 4 Argon (input $2); the gap is within the margin of error, so Gemini 4 Argon is enough.
Rating and price
The closer to the top left, the cheaper and better rated the model.
OpenAIAnthropicZ.aiGoogleAlibabaMistralCohereMetaMiniMaxxAIDeepSeekMoonshotNVIDIAMicrosoftOther
Best-rated model at each price
First line: model / Human rating, second line: price estimate
| Model | Developer | Human rating | Price |
|---|---|---|---|
| △Claude Opus 5.5 | Anthropic | In $4 (out $20) | |
| Gemini 4 Argon | In $2 (out $10) | ||
| Gemini 3.8 Flash | In $0.75 (out $3.75) | ||
| GLM-5.3 Flash | Z.ai | In $0.15 (out $0.5) | |
| GPT-6 Luna | OpenAI | In $0.1 (out $0.5) | |
| GPT-5 nano | OpenAI | In $0.05 (out $0.4) |
All models (highest human rating first)
Second line: price (input / output) / context / released / supports Tools Image JSON □ Compare
| Pick | Compare | Model | Developer | Human rating | Japanese (human rating) | JA instructions (%) | Input $ /1M | Output $ /1M | Context | Supports | Weights | Released | Price and specs |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Claude Opus 5.5 | Anthropic | 1532 | — | $4 | $20 | 1M | $4 / $20 1M 2026-09 | ||||||
| Gemini 4 Argon | — | — | $2 | $10 | 1.05M | $2 / $10 1.05M 2026-09 | |||||||
| Claude Opus 4.6 | Anthropic | 1499 | — | $5 | $25 | 1M | $5 / $25 1M 2026-02 | ||||||
| Claude Fable 5 | Anthropic | 1521 | — | $10 | $50 | 1M | $10 / $50 1M 2026-06 | ||||||
| Claude Opus 4.7 | Anthropic | 1490 | — | $5 | $25 | 1M | $5 / $25 1M 2026-04 | ||||||
| Claude Fable 5.1 | Anthropic | 1551 | — | $10 | $50 | 1M | $10 / $50 1M 2026-09 | ||||||
| Muse Wonder | Meta | — | — | $4 | $20 | 1.05M | $4 / $20 1.05M 2026-10 | ||||||
| Claude Opus 5 | Anthropic | 1509 | — | $5 | $25 | 1M | $5 / $25 1M 2026-07 | ||||||
| Claude Opus 4.8 | Anthropic | 1461 | — | $5 | $25 | 1M | $5 / $25 1M 2026-05 | ||||||
| GPT-5.6 Sol | OpenAI | 1497 | — | $4 | $20 | 1.05M | $4 / $20 1.05M 2026-07 | ||||||
| Gemini 3.8 Flash | 1508 | — | $0.75 | $3.75 | 1.05M | $0.75 / $3.75 1.05M 2026-09 | |||||||
| Kimi K3 | Moonshot | 1511 | — | $3 | $15 | 1.05M | $3 / $15 1.05M 2026-07 | ||||||
| Gemini 3.7 Flash | 1513 | — | $0.75 | $3.75 | 1.05M | $0.75 / $3.75 1.05M 2026-08 | |||||||
| Claude Sonnet 5.5 | Anthropic | 1468 | — | $2 | $10 | 1M | $2 / $10 1M 2026-09 | ||||||
| GPT-6.1 Sol | OpenAI | — | — | $2 | $10 | 1.05M | $2 / $10 1.05M 2026-09 | ||||||
| Muse Spark 1.3 | Meta | 1476 | — | $1.25 | $4.25 | 1.05M | $1.25 / $4.25 1.05M 2026-09 | ||||||
| Claude Opus 4.5 | Anthropic | 1447 | — | $5 | $25 | 200K | $5 / $25 200K 2025-11 | ||||||
| Gemini 3.1 Pro (preview) | 1496 | — | $2 | $12 | 1.05M | $2 / $12 1.05M 2026-02 | |||||||
| GLM-5.3 | Z.ai | 1467 | — | $1.4 | $4.4 | 1M | $1.4 / $4.4 1M 2026-08 | ||||||
| GPT-5.5 | OpenAI | 1505 | — | $5 | $30 | 1.05M | $5 / $30 1.05M 2026-04 | ||||||
| DeepSeek-V4.1 Flash | DeepSeek | 1453 | — | — | — | 1M | 1M 2026-09 | ||||||
| Qwen3.8 Max | Alibaba | 1478 | — | $2 | $6 | 1M | $2 / $6 1M 2026-08 | ||||||
| Claude Sonnet 4.6 | Anthropic | 1441 | — | $3 | $15 | 1M | $3 / $15 1M 2026-02 | ||||||
| GPT-6 Astra | OpenAI | 1437 | — | $10 | $50 | 1.05M | $10 / $50 1.05M 2026-09 | ||||||
| GLM-5.3 Flash | Z.ai | 1450 | — | $0.15 | $0.5 | 1M | $0.15 / $0.5 1M 2026-08 | ||||||
| Gemini 3.6 Flash | 1476 | — | $0.75 | $3.75 | 1.05M | $0.75 / $3.75 1.05M 2026-07 | |||||||
| Gemini 3 Pro | 1504 | — | — | — | 1.05M | 1.05M 2025-11 | |||||||
| Muse Spark 1.2 | Meta | — | — | $1.25 | $4.25 | 1.05M | $1.25 / $4.25 1.05M 2026-08 | ||||||
| Muse Spark 1.1 | Meta | 1468 | — | $1.25 | $4.25 | 1M | $1.25 / $4.25 1M 2026-07 | ||||||
| MiMo-V2.6 Pro | Xiaomi | — | — | $0.435 | $0.87 | 1.05M | $0.435 / $0.87 1.05M 2026-09 | ||||||
| GPT-5.4 | OpenAI | 1480 | 88.1 | $2.5 | $15 | 1.05M | $2.5 / $15 1.05M 2026-03 | ||||||
| GLM-5.2 | Z.ai | 1454 | — | $1.4 | $4.4 | 1M | $1.4 / $4.4 1M 2026-06 | ||||||
| MiMo-V2.5 Pro | Xiaomi | 1420 | — | $0.435 | $0.87 | 1.05M | $0.435 / $0.87 1.05M 2026-04 | ||||||
| Claude Sonnet 5 | Anthropic | 1443 | — | $2 | $10 | 1M | $2 / $10 1M 2026-06 | ||||||
| Gemini 3.5 Flash | 1479 | — | $1.5 | $9 | 1.05M | $1.5 / $9 1.05M 2026-05 | |||||||
| Claude Sonnet 4.5 | Anthropic | 1413 | — | $3 | $15 | 200K | $3 / $15 200K 2025-09 | ||||||
| GPT-5.6 Terra | OpenAI | 1471 | — | $2 | $12 | 1.05M | $2 / $12 1.05M 2026-07 | ||||||
| GLM-5.1 | Z.ai | 1436 | — | $1.4 | $4.4 | 200K | $1.4 / $4.4 200K 2026-04 | ||||||
| Grok 4.5 | xAI | 1444 | — | $2 | $6 | 500K | $2 / $6 500K 2026-07 | ||||||
| DeepSeek-V4 Pro (2026-08) | DeepSeek | — | — | $0.66 | $1.98 | 1M | $0.66 / $1.98 1M 2026-08 |
How to read this page
- Also shown (not used for ranking): human ratings on questions in Japanese; success rates on Japanese instruction following. Prices are compared by the input price. 392 models have ratings (157 match the current conditions).
- Bars are drawn from 1150 to 1550.
- The human rating comes from people giving two models the same question and picking the better answer. With a gap of 50, the higher-rated model is picked about 57% of the time; 64% at 100 and 76% at 200 (a gap of 0 is even odds). Differences where the ± ranges overlap are within the margin of error, so about the same.
- Only some models have values for JA instructions, mostly models with published weights.
- Only models rated higher than every cheaper model are listed, best first. For about the same rating, pick the cheapest row in this table. The top row without △ within the price you can pay is the pick. A △ row costs more than the row below it without △, yet its rating is about the same (the gap is within the margin of error).
- “On a budget” is the highest-rated model at $0.5 or less per million input tokens (or at your price cap, when that is lower).
- △ marks rows whose rating gap is within the margin of error (about the same). A row without △ is rated higher than the nearest row below it without △, by more than the margin of error.
- Best at each price
- Prices are what developers pay to use a model through its API, separate from app subscriptions. The unit is one million tokens. A token is a small piece of text (about three quarters of an English word); one million tokens is roughly 8 to 10 novels. Here this is the input price. For example, at $4, one everyday question and answer (about 1,000 tokens) costs about 0.4 cents. Models without a known price are left out.
- 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.
- Vertical is Human rating, horizontal is price (the horizontal scale gets coarser to the right). Points joined by the step line are the best-rated models at each price. Hover or tap a dot to see the model's name.
Data: Arena, Epoch AI, Swallow LLM team (CC BY 4.0), Models.dev (MIT), updated 2026-10-24, data sources