> ## Documentation Index > Fetch the complete documentation index at: https://docs.priorlabs.ai/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > TabPFN-3 and later versions support up to 1,000,000 rows, subject to feature count and checkpoint/API limits. > For large datasets, use per-estimator subsampling, e.g. inference_config={"SUBSAMPLE_SAMPLES": 100_000}, in TabPFNClassifier or TabPFNRegressor from tabpfn or tabpfn_client. > Pass raw pandas DataFrames to fit and predict. Categorical strings/categories, raw text, and missing feature values are handled automatically; no manual integer/one-hot encoding, imputation, scaling, or outlier removal is needed. # TabPFN-3.5 > What's new in TabPFN-3.5 — text, high-cardinality and wide-table support, one multitask checkpoint, Fast, Plus and Thinking, and migration guidance. TabPFN-3.5 improves most on real-world prediction tasks, where the incoming ERP, CRM and production data isn’t clean, with the model family ranking first on TabArena and BeyondArena. [Plus](#tabpfn-client-and-api) and [Thinking](#tabpfn-client-and-api) are updated to version 3.5 as well and deliver additional gains on grouped, temporal and text-rich data. We also introduce an alpha fast inference model, [TabPFN-3.5-Fast](/capabilities/fast-checkpoint). ## Model TabPFN-3.5 is our first multitask model, improving both classification and regression within a single checkpoint. It achieves state-of-the-art performance on BeyondArena’s wide datasets and improves on high-cardinality categories, reducing the need for manual encoding and feature selection. With that, we also increase the [maximum number of features supported](/models#tabpfn-3-5-family) tenfold to 20,000, with up to 6,000 recommended for better performance. TabPFN-3.5 also improves [predictive distributions](/capabilities/predictive-distribution) for skewed and zero-heavy regression targets such as claim amounts and customer spend. It ranks first on ScoringBench, which evaluates the quality of predictive distributions. Building on the base model, [TabPFN-3.5-Plus](#tabpfn-client-and-api) significantly improves performance on text-rich data, making it well-suited for tasks where signal comes from product descriptions, reviews or inspection notes. [TabPFN-3.5-Thinking](#tabpfn-client-and-api), our most accurate model, further improves on grouped and temporal data. This helps with tasks where rows belong to related entities or observations follow a time sequence. Applications include fraud scoring across accounts, churn prediction from customer histories, and demand forecasting across stores and products. [TabPFN-3.5-Fast](/capabilities/fast-checkpoint), available in alpha, delivers up to 6× faster inference than the base model for latency-sensitive applications. See the [TabPFN-3.5 model family](/models#tabpfn-3-5-family) page and our [technical report](https://priorlabs.ai/technical-reports/tabpfn-3-5) for more details. *** ## Open-source package TabPFN-3.5 and [TabPFN-3.5-Fast](/capabilities/fast-checkpoint) (alpha) model weights are available for local inference and research starting today. The package downloads the weights after a one-time license acceptance. [Check the access guide](/models/accessing-model-weights) for setup instructions and [model licensing](/models#tabpfn-model-license) for usage terms. The `9.0.0` release of the `tabpfn` open-source package keeps the existing interfaces. This release also adds: * **Built-in text and datetime preprocessing.** Local inference can process [text columns](/capabilities/text-features) and [datetime columns](/improving-performance/preprocessing#datetime-features) directly into features, reducing the time spent on feature engineering. * **Inference defaults.** To improve performance, we have introduced model-specific inference configurations. With that, `softmax_temperature` now defaults to `"auto"`. Both it and `n_estimators` read their defaults from the selected checkpoint. * **Keep rare outcomes when subsampling.** The new [`majority_downsample`](/improving-performance/preprocessing#row-subsampling) option keeps all non-majority rows while sampling the most frequent target value, including zeros in regression. * **Smaller KV caches for `TabPFNClassifier`.** Classification caches now store projected decoder keys instead of full embeddings. See [KV cache](/capabilities/kv-cache). * **More efficient processing of large numeric tables.** We optimized RAM and GPU memory across preprocessing, `fit()` and `predict()`, with 41% lower peak RAM in our large-table benchmark. We also shipped: * [Categorical feature declarations](/improving-performance/preprocessing#categorical-features) in string columns are now respected regardless of how many distinct values a column contains. The declaration can now also be made through the `pandas` `category` dtype. * Saved models now load with `device="auto"` by default ([#1243](https://github.com/PriorLabs/TabPFN/pull/1243)). * [Fine-tuning](/capabilities/fine-tuning) now chunks large batches to prevent a memory-access crash ([#1184](https://github.com/PriorLabs/TabPFN/pull/1184)). * Classes absent from the validation set now receive neutral thresholds during [metric tuning](/capabilities/metric-tuning) ([#1152](https://github.com/PriorLabs/TabPFN/pull/1152)). See [Migrating from TabPFN-3](#migrating-from-tabpfn-3) for behavior changes. *** ## `tabpfn-client` and API [TabPFN-3.5-Plus](/models#tabpfn-3-5-family) improves results on text-rich tables by adding [proprietary text processing](/capabilities/text-features) to the base model. It ranks #1 on STRABLE, a benchmark for tabular predictions over tables with strings, and is available through the [hosted API](/api-reference/getting-started), [`tabpfn-client`](/quickstart), [MCP](/agentic/mcp), and [integrations](#integrations). [TabPFN-3.5-Thinking](/capabilities/thinking-mode) is our most accurate offering, especially on grouped and temporal data. It uses more compute during fitting to improve prediction quality, and the fitted model can then be reused for repeated predictions. Thinking is available through the hosted API, `tabpfn-client`, and [enterprise deployments](#integrations); see [Models](/models#tabpfn-3-5-family) for its limits. [TabPFN-3.5-Fast](/capabilities/fast-checkpoint) (alpha) and Base models are also available through the API and can be selected through the TabPFN instance constructor. We have also updated [API metering](/api-reference/metering), which now accounts for the selected model and operation performed, including 75% lower charges for successful KV cache reuse. Thinking fits and predictions share the same token budgets, replacing the separate Thinking fit limit. From 15 September until 29 September 2026, all API and MCP users get a 50% reduction from the standard TabPFN-3.5 token rates for the 3.5 family. Standard rates resume on 29 September. The `0.6.0` release of `tabpfn-client` adds: * **Estimate token cost before running a request.** `estimate_cost()` sends only dataset dimensions and settings and returns cost in tokens, without uploading data or consuming usage limits. See [API metering](/api-reference/metering#estimate-and-monitor-usage) for more details. * **Reuse fitted models across Python sessions.** [`save_model()` and `load_model()`](https://github.com/PriorLabs/tabpfn-client#saving-and-loading-fitted-models) let you resume predictions without fitting again. The saved record references the server-side model; use the same account. See Thinking mode and [KV cache](/capabilities/kv-cache) for more details. * **Inspect prediction uncertainty visually.** [`plot_regression_distribution()`](https://github.com/PriorLabs/tabpfn-client#plotting) plots the predicted regression distribution. Install `tabpfn-client[viz]` to use it. We also shipped: * `fit()` and `predict()` no longer print progress spinners, and routine messages use debug logging ([#385](https://github.com/PriorLabs/tabpfn-client/pull/385)). * Restored compatibility with `scikit-learn` tools such as partial dependence ([#372](https://github.com/PriorLabs/tabpfn-client/pull/372)). *** ## Integrations SAP AI Core is a new way to access [TabPFN-3.5-Plus](/models#tabpfn-3-5-family), available today to all SAP customers. TabPFN-3.5-Plus and [Fast](/capabilities/fast-checkpoint) (alpha) are available on [AWS SageMaker](/integrations/sagemaker), with [Thinking](/capabilities/thinking-mode) available through a private listing. TabPFN-3.5-Plus is also available on [Microsoft Foundry](/integrations/foundry). The TabPFN-3.5 model family is also available as an on-premises container. For enterprise upgrades, container access or private listings, contact [our sales team](http://priorlabs.ai/contact). *** ## Migrating from TabPFN-3 For both the local package and API: 1. Compare TabPFN-3.5 with TabPFN-3 on your validation set. Test raw text and categorical columns before removing existing encoders or feature selection. See [Preprocessing](/improving-performance/preprocessing). 2. Declare categorical identifiers explicitly so they are not mistaken for numerical or free-text features. See [Feature engineering](/improving-performance/feature-engineering). 3. Fit a new model when switching versions. Existing fitted models and [KV caches](/capabilities/kv-cache) remain tied to the version used to create them. 4. Check the [model limits](/models#tabpfn-3-5-family) for your selected variant, as well as for text feature limits. ### Open-source package 1. Upgrade with `pip install --upgrade tabpfn` and complete the [weight-access setup](/models/accessing-model-weights). 2. Select `ModelVersion.V3_5` for the base model or `ModelVersion.V3_5_FAST` for [Fast](/capabilities/fast-checkpoint) through `create_default_for_version()`. See [Selecting Model Version](/models/selecting-model-version#oss-package). 3. For [built-in text processing](/improving-performance/preprocessing#text-features), use `pandas` `string` or `pyarrow` string columns. Undeclared string columns with more than 30 distinct values qualify as text. 4. Keep dates in a [datetime dtype](/improving-performance/preprocessing#datetime-features) for them to be processed correctly. 5. Keep the selected model’s [defaults](/improving-performance/preprocessing#what-happens-by-default) when customizing inference. Pass only your changes in an `inference_config` dictionary. Passing a full `InferenceConfig` object is still supported but is deprecated. 6. If you extract training embeddings, refit with `fit_mode="fit_preprocessors"`. Training embeddings are no longer available with `fit_with_cache`; test embeddings remain available. 7. Existing fine-tuned checkpoints remain based on TabPFN-3. To use the new model, repeat [fine-tuning](/capabilities/fine-tuning) with TabPFN-3.5. Fine-tuning estimators still require text and datetime features to be converted beforehand. ### API client 1. Upgrade with `pip install --upgrade tabpfn-client`. 2. Select [Plus](/models#tabpfn-3-5-family) with `create_default_for_version("v3.5")`, or Fast with `create_default_for_version("v3.5-fast")`. Use `"v3"` to keep using TabPFN-3-Plus. See [Selecting Model Version](/models/selecting-model-version#python-client). 3. For [TabPFN-3.5-Thinking](/capabilities/thinking-mode), `thinking_mode=True` selects the 3.5 version. If you'd like to use TabPFN-3-Thinking, pass `"v3"` as the model version. 4. Review the updated [API metering](/api-reference/metering). Token charges depend on the selected model and settings. ### REST API 1. Select `model_path="v3.5_default"` or `"v3.5-fast_default"` inside `tabpfn_config` - see [Selecting Model Version](/models/selecting-model-version#rest-api) for more details. 2. The upload, fit, and predict flow is unchanged. 3. Check `GET /tabpfn/get_model_limits` for the API's current limits. ### Enterprise deployments Deploy the updated image or marketplace model to use TabPFN-3.5. Existing endpoints do not upgrade automatically. Follow the [SageMaker](/integrations/sagemaker) or [Foundry](/integrations/foundry) deployment guide for your platform. *** ## Get started Compare capabilities, limits, and availability. Configure fit-time optimization for better predictions. Reuse training-set computation for repeated predictions. Prepare your data and tune prediction quality.