# Fine-tuning services for AI models

> 6 fine-tuning listings ranked by the Anchor benchmark. Leader Vertex AI Gemini tuning (B). Services that train a model on your examples, by supervised, preference or reinforcement fine-tuning, and serve the result. Compared on which base models you can tune, the methods, price per training token, whether you get the weights and what serving the result costs.

- Canonical: https://www.anchorterminal.com/categories/fine-tuning
- Markdown: https://www.anchorterminal.com/categories/fine-tuning.md (~1,900 tokens)
- Slim: https://www.anchorterminal.com/categories/fine-tuning.min.md (~430 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/categories/fine-tuning.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- Updated: 2026-10-04

Services that train a model on your examples, by supervised, preference or reinforcement fine-tuning, and serve the result. Compared on which base models you can tune, the methods, price per training token, whether you get the weights and what serving the result costs.

- Tools ranked: 6 · agent-ready (BB or better): 0 · accept x402: 0 · hosted endpoints: 4 · desk reviews by the panel: 12
- JSON: https://www.anchorterminal.com/api/v1/tools.json (list) · https://www.anchorterminal.com/api/v1/rankings.json (ranked) · https://www.anchorterminal.com/api/v1/x402.json (payable) · https://www.anchorterminal.com/api/v1/capabilities.json (by capability)
- Grades run AA, A, BB, B, C, D, E, F · methodology: https://www.anchorterminal.com/benchmark/

- Capabilities in this category: finetune.sft, finetune.preference, finetune.rl, finetune.lora, finetune.export
- https://letme.dev/finetune.sft picks the top-graded tool in this list and says how to call it direct; calling through letme comes later (https://www.anchorterminal.com/letme/index.md)

## Ranking

| # | Tool | Vendor | Kind | Category | Grade | Score | Confidence | x402 | Auth | Where | Reviews | Page |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 190 | Vertex AI Gemini tuning | Google Cloud | HTTP API | Fine-tuning | B | 64.2 | medium | no | OAuth | hosted | 2.5/5 (2) | https://www.anchorterminal.com/tools/vertex-ai-tuning.md |
| 228 | Microsoft Foundry fine-tuning (Azure OpenAI) | Microsoft Azure | HTTP API | Fine-tuning | C | 61.4 | medium | no | OAuth or key | hosted | 3.5/5 (2) | https://www.anchorterminal.com/tools/azure-foundry-fine-tuning.md |
| 269 | Fireworks AI Fine-tuning | Fireworks AI | HTTP API | Fine-tuning | C | 59.2 | medium | no | API key | hosted | 2.5/5 (2) | https://www.anchorterminal.com/tools/fireworks-fine-tuning.md |
| 319 | Together AI Fine-tuning | Together AI | HTTP API | Fine-tuning | C | 54.9 | medium | no | API key | hosted | 3/5 (2) | https://www.anchorterminal.com/tools/together-fine-tuning.md |
| 347 | Unsloth | Unsloth | Agent framework | Fine-tuning | D | 51.7 | medium | no | None | library | 3.5/5 (2) | https://www.anchorterminal.com/tools/unsloth.md |
| 354 | Tinker | Thinking Machines Lab | SDK + MCP | Fine-tuning | D | 51.2 | medium | no | API key | local | 3.5/5 (2) | https://www.anchorterminal.com/tools/tinker.md |

Scores are from public evidence against the published checklist (https://www.anchorterminal.com/benchmark/), with Performance and Task success pending. p95 latency and context cost come from our probes, which haven't run yet.

## Summaries

### 190. Vertex AI Gemini tuning, B (64.2)

Supervised, preference and reinforcement tuning of Gemini, plus supervised tuning of Gemma, Llama and Qwen, on Google Cloud's Gemini Enterprise Agent Platform (the platform formerly called Vertex AI). Supervised, preference and reinforcement tuning of Gemini, plus supervised tuning of Gemma, Llama and Qwen. No weight export. The tuned model exists only as a Google Cloud endpoint.

- Page: https://www.anchorterminal.com/tools/vertex-ai-tuning · Markdown: https://www.anchorterminal.com/tools/vertex-ai-tuning.md · JSON: https://www.anchorterminal.com/api/v1/tools/vertex-ai-tuning.json
- Capabilities: finetune.sft, finetune.preference, finetune.rl, finetune.lora · endpoint: `https://us-central1-aiplatform.googleapis.com/v1`

### 228. Microsoft Foundry fine-tuning (Azure OpenAI), C (61.4)

Azure's managed service for supervised, preference and reinforcement fine-tuning of supported OpenAI and open-weight models. SFT, DPO and RFT on GPT-4.1 and o4-mini through the OpenAI-shaped /openai/v1 API. No weight export; checkpoints copy only between Azure resources.

- Page: https://www.anchorterminal.com/tools/azure-foundry-fine-tuning · Markdown: https://www.anchorterminal.com/tools/azure-foundry-fine-tuning.md · JSON: https://www.anchorterminal.com/api/v1/tools/azure-foundry-fine-tuning.json
- Capabilities: finetune.sft, finetune.preference, finetune.rl, finetune.lora · endpoint: `https://<resource>.openai.azure.com/openai/v1`

### 269. Fireworks AI Fine-tuning, C (59.2)

Managed supervised, preference and reinforcement fine-tuning for open models, with a training API for custom workflows. SFT, DPO, ORPO and RFT as managed jobs, plus a serverless Training API that is generally available. Tuned LoRAs only deploy to on-demand GPUs at $8 an hour and up, never to serverless.

- Page: https://www.anchorterminal.com/tools/fireworks-fine-tuning · Markdown: https://www.anchorterminal.com/tools/fireworks-fine-tuning.md · JSON: https://www.anchorterminal.com/api/v1/tools/fireworks-fine-tuning.json
- Capabilities: finetune.sft, finetune.preference, finetune.rl, finetune.lora, finetune.export · endpoint: `https://api.fireworks.ai`

### 319. Together AI Fine-tuning, C (54.9)

Managed LoRA and full fine-tuning, supervised or DPO, on about 30 open models from Qwen3.5 0.8B to Kimi K2.7, billed per training token with a $4 minimum. 31 tunable base models, 11 or 12 of them with full fine-tuning as well as LoRA. Fine-tuned models don't run serverless; dedicated endpoints start at $5.49 an hour.

- Page: https://www.anchorterminal.com/tools/together-fine-tuning · Markdown: https://www.anchorterminal.com/tools/together-fine-tuning.md · JSON: https://www.anchorterminal.com/api/v1/tools/together-fine-tuning.json
- Capabilities: finetune.sft, finetune.preference, finetune.lora, finetune.export · endpoint: `https://api.together.ai/v1`

### 347. Unsloth, D (51.7)

Open-source library, web UI (Studio) and desktop app for LoRA, QLoRA, full fine-tuning and RL (GRPO, DPO, ORPO) of open models on your own GPU, from 3 GB of VRAM. The Apache-2.0 core runs on customer hardware and keeps model weights there. Users supply and pay for the GPU.

- Page: https://www.anchorterminal.com/tools/unsloth · Markdown: https://www.anchorterminal.com/tools/unsloth.md · JSON: https://www.anchorterminal.com/api/v1/tools/unsloth.json
- Capabilities: finetune.sft, finetune.preference, finetune.rl, finetune.lora, finetune.export

### 354. Tinker, D (51.2)

Thinking Machines Lab's API for model training. Full control of the training loop with the GPUs abstracted away, plus recipes for SFT, DPO, RL and distillation. LoRA only; no full-parameter training.

- Page: https://www.anchorterminal.com/tools/tinker · Markdown: https://www.anchorterminal.com/tools/tinker.md · JSON: https://www.anchorterminal.com/api/v1/tools/tinker.json
- Capabilities: finetune.sft, finetune.preference, finetune.rl, finetune.lora, finetune.export

## How we test this category

The same small dataset used to tune a comparable open model on each service, then served. We check the job flow, how long training takes, whether the weights can leave, and the training and serving cost. This test hasn't run yet, so Task success is pending and the grades here come from the categories assessed from public evidence.