> ## Documentation Index
> Fetch the complete documentation index at: https://pioneer-kelton-add-decoder-inference-prices.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Pioneer model catalog: encoders, decoders, and inference

> Browse Pioneer's encoder (GLiNER) and decoder (LLM) models for fine-tuning and inference. Covers on-demand vs. serverless and how to query the live catalog.

Pioneer supports two model families: **encoder models** (GLiNER) for structured extraction tasks like named entity recognition, and **decoder models** (LLMs) for text generation, classification, and open-ended prompting. The tables below are a snapshot of the current catalog — use `GET /base-models` to query the live list, which always reflects current availability and capabilities.

Some rollout-stage models are feature-gated. They appear in the live catalog only for workspaces that have the corresponding rollout enabled.

## Encoder models (GLiNER)

GLiNER models perform named entity recognition and structured extraction. Most GLiNER base models support both training and on-demand inference after training. Prices are per 1M tokens.

| Model ID                                   | Label                              | Input  | Output | Training   | Inference             |
| :----------------------------------------- | :--------------------------------- | :----- | :----- | :--------- | :-------------------- |
| `fastino/gliner2-base-v1`                  | GLiNER2 Base                       | \$0.15 | \$0.15 | LoRA, Full | Serverless, On-demand |
| `fastino/gliner2-large-v1`                 | GLiNER2 Large                      | \$0.15 | \$0.15 | LoRA, Full | Serverless, On-demand |
| `fastino/gliner2-multi-v1`                 | GLiNER2 Multi                      | \$0.15 | \$0.15 | LoRA, Full | Serverless, On-demand |
| `fastino/gliner2-multi-large-v1`           | GLiNER2 Multi Large                | \$0.15 | \$0.15 | LoRA, Full | Serverless, On-demand |
| `fastino/gliguard-LLMGuardrails-300M`      | GLiGuard LLM Guardrails 300M       | \$0.15 | \$0.15 | —          | Serverless            |
| `fastino/gliner2-privacy-filter-PII-multi` | GLiNER2 Privacy Filter PII (Multi) | \$0.15 | \$0.15 | —          | Serverless            |
| `fastino/gliguard-PII-multi`               | GLiNER2-Guardrails-PII-Multi       | \$0.15 | \$0.15 | —          | Serverless            |

`fastino/gliner2-multi-v1` and `fastino/gliner2-multi-large-v1` are multilingual variants suitable for non-English text.

## Decoder models — training

LoRA fine-tuning via `POST /felix/training-jobs` is limited to the Nemotron 3.5 Lightning family:

| Model ID                                            | Label                                     | Training |
| :-------------------------------------------------- | :---------------------------------------- | :------- |
| `nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16` | Nemotron 3.5 Lightning 30B-A3B            | LoRA     |
| `fastino/Fastino-Nemotron-3.5-Lightning-Finance`    | Fastino Nemotron 3.5 Lightning Finance    | LoRA     |
| `fastino/Fastino-Nemotron-3.5-Lightning-Healthcare` | Fastino Nemotron 3.5 Lightning Healthcare | LoRA     |

Together with the trainable GLiNER2 Base, Large, Multi, and Multi Large encoder
targets above, these are the only supported base models for new training jobs. Query
`GET /base-models?supports_training=true` before submitting a job; it is the
live source of truth for availability.

## Decoder models — serverless inference

These are the supported pre-deployed inference families. Rates are USD per 1M tokens.
Query `GET /base-models?supports_inference=true` for live availability, context limits,
and pricing — it is the source of truth if it ever disagrees with this page.

<Note>
  **Cached input** is what a prompt-cache hit costs (see [prompt caching](#prompt-caching)).
  Where it equals the input rate, that route does not price cache reads separately.
  A dash means the rate is not published for that model — rollout-gated models return
  their rates only to workspaces with the rollout enabled.
</Note>

### Nemotron 3.5 Lightning

| Model ID                                            | Label                                     | Input  | Output | Cached input | Context |
| --------------------------------------------------- | ----------------------------------------- | ------ | ------ | ------------ | ------- |
| `nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16` | Nemotron 3.5 Lightning 30B-A3B            | \$0.50 | \$0.50 | \$0.50       | 8,192   |
| `fastino/Fastino-Nemotron-3.5-Lightning-Finance`    | Fastino Nemotron 3.5 Lightning Finance    | \$0.50 | \$0.50 | \$0.50       | 8,192   |
| `fastino/Fastino-Nemotron-3.5-Lightning-Healthcare` | Fastino Nemotron 3.5 Lightning Healthcare | \$0.50 | \$0.50 | \$0.50       | 8,192   |

### Anthropic

| Model ID             | Label                | Input   | Output  | Cached input | Context   |
| -------------------- | -------------------- | ------- | ------- | ------------ | --------- |
| `claude-opus-5`      | Claude Opus 5        | \$5.00  | \$25.00 | \$0.50       | 1,000,000 |
| `claude-sonnet-5`    | Claude Sonnet 5      | \$2.00  | \$10.00 | \$0.20       | 1,000,000 |
| `claude-opus-5-fast` | Claude Opus 5 (Fast) | \$10.00 | \$50.00 | \$1.00       | 1,000,000 |
| `claude-fable-5`     | Claude Fable 5       | \$11.00 | \$55.00 | \$1.10       | 1,000,000 |
| `claude-haiku-4-5`   | Claude Haiku 4.5     | \$1.00  | \$5.00  | \$0.10       | 200,000   |

### OpenAI

| Model ID        | Label         | Input  | Output  | Cached input | Context   |
| --------------- | ------------- | ------ | ------- | ------------ | --------- |
| `gpt-5.5`       | GPT-5.5       | \$5.00 | \$30.00 | \$0.50       | 1,000,000 |
| `gpt-5.6-luna`  | GPT-5.6 Luna  | \$1.00 | \$6.00  | \$0.10       | 1,050,000 |
| `gpt-5.6-terra` | GPT-5.6 Terra | \$2.50 | \$15.00 | \$0.25       | 1,050,000 |
| `gpt-5.6-sol`   | GPT-5.6 Sol   | \$5.00 | \$30.00 | \$0.50       | 1,050,000 |

<Warning>
  **GPT-5 models bill a long-context tier.** Requests whose prompt exceeds
  **272,000 tokens** are charged at roughly double the rates above — GPT-5.5 and
  GPT-5.6 Sol go to \$10.00 / \$45.00 / \$1.00, GPT-5.6 Terra to \$5.00 / \$22.50 /
  \$0.50, and GPT-5.6 Luna to \$2.00 / \$9.00 / \$0.20. The rest of this page's
  rates are the standard tier.
</Warning>

### DeepSeek

| Model ID                        | Label             | Input  | Output | Cached input | Context   |
| ------------------------------- | ----------------- | ------ | ------ | ------------ | --------- |
| `deepseek-ai/DeepSeek-V4-Flash` | DeepSeek V4 Flash | \$0.14 | \$0.28 | \$0.028      | 1,000,000 |

The dated pin `deepseek-ai/DeepSeek-V4-Flash-0731` remains a working alias for this
entry — both resolve to the same 0731 checkpoint upstream.

### Z.ai

| Model ID          | Label   | Input  | Output | Cached input | Context   |
| ----------------- | ------- | ------ | ------ | ------------ | --------- |
| `zai-org/GLM-5.2` | GLM 5.2 | \$1.40 | \$4.40 | \$0.14       | 1,040,000 |

### Moonshot AI

| Model ID                  | Label        | Input  | Output  | Cached input | Context   |
| ------------------------- | ------------ | ------ | ------- | ------------ | --------- |
| `moonshotai/Kimi-K3`      | Kimi K3      | \$3.00 | \$15.00 | \$0.30       | 1,000,000 |
| `moonshotai/Kimi-K3-Fast` | Kimi K3 Fast | \$4.50 | \$22.50 | \$0.45       | 1,000,000 |

<Note>
  Models outside these families are not supported inference targets. Use
  `GET /base-models` for the live catalog and lifecycle state before integrating.
</Note>

## Prompt caching

Many serverless models bill cached input tokens at a discount, and some providers bill a one-time surcharge to write tokens into the cache. Pioneer passes each provider's published cache rates straight through — these are the same rates `GET /base-models` returns as `cache_read_price_per_million` and `cache_write_price_per_million`, and the same rates Pioneer bills you.

Cache rates are derived from each model's input rate using the multipliers below. Where a provider has no separate cache-write line item, cache writes bill at the standard input rate.

| Provider              | Cache read | Cache write                                  |
| --------------------- | ---------- | -------------------------------------------- |
| Anthropic (Claude)    | 0.1× input | 1.25× input                                  |
| OpenAI (GPT-5 family) | 0.1× input | input rate (1.25× on GPT-5.6 Luna/Sol/Terra) |

Query the live catalog for exact cache rates. Any model without an explicit
cache discount bills cached input at the standard input rate.

## On-demand vs. serverless inference

Pioneer offers two ways to serve predictions, and the right choice depends on your workflow.

**Serverless** inference uses pre-deployed base model endpoints. There is no startup delay and you are billed per token. This is ideal when you want to call a frontier model without fine-tuning.

**On-demand** inference provisions a dedicated GPU after fine-tuning completes. Your LoRA adapter is loaded onto the GPU and served exclusively for your requests. Pioneer routes inference calls to an on-demand deployment automatically when you pass a training job ID as `model_id`.

## Querying the live catalog

The tables above may lag behind newly added models. Use `GET /base-models` to get the current catalog at runtime.

```bash theme={null}
# All models
curl https://api.pioneer.ai/base-models \
  -H "X-API-Key: YOUR_API_KEY"

# Only models that support inference
curl "https://api.pioneer.ai/base-models?supports_inference=true" \
  -H "X-API-Key: YOUR_API_KEY"

# Only models that support training
curl "https://api.pioneer.ai/base-models?supports_training=true" \
  -H "X-API-Key: YOUR_API_KEY"

# Filter by model family
curl "https://api.pioneer.ai/base-models?task_type=encoder" \
  -H "X-API-Key: YOUR_API_KEY"

curl "https://api.pioneer.ai/base-models?task_type=decoder" \
  -H "X-API-Key: YOUR_API_KEY"
```

Each entry in the response includes the model ID, its display label, context length, per-1M-token rates (`input_price_per_million`, `output_price_per_million`, `cache_read_price_per_million`, `cache_write_price_per_million`), and boolean flags for `supports_training` and `supports_inference`. Use the model ID value directly in training job requests and inference calls.

To list every model alongside its input, output, and cache rates:

```bash theme={null}
curl -s "https://api.pioneer.ai/base-models?supports_inference=true" \
  -H "X-API-Key: YOUR_API_KEY" \
| jq -r '["model","input/M","output/M","cache_read/M","cache_write/M"],
    (.models[] | [.id, .input_price_per_million, .output_price_per_million,
      .cache_read_price_per_million, .cache_write_price_per_million]) | @tsv' \
| column -t -s $'\t'
```