# Tinker vs Unsloth

> Unsloth has a score of 51.7 (D) against Tinker's 51.2 (D). Both do finetune sft. The largest gap is payments & pricing, 40 points. Category scores, facts, verdicts and agent notes side by side.

- Canonical: https://www.anchorterminal.com/compare/tinker-vs-unsloth
- Markdown: https://www.anchorterminal.com/compare/tinker-vs-unsloth.md (~1,350 tokens)
- Slim: https://www.anchorterminal.com/compare/tinker-vs-unsloth.min.md (~330 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/compare/tinker-vs-unsloth.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: 2026-10-05

Unsloth has a score of 51.7 (D) against Tinker's 51.2 (D). Both do finetune sft. The largest gap is payments & pricing, 40 points.

- Tinker: grade D, 51.2/100, rank #354 of 452. Markdown https://www.anchorterminal.com/tools/tinker.md · JSON https://www.anchorterminal.com/api/v1/tools/tinker.json
- Unsloth: grade D, 51.7/100, rank #347 of 452. Markdown https://www.anchorterminal.com/tools/unsloth.md · JSON https://www.anchorterminal.com/api/v1/tools/unsloth.json

## Which one, for what

Pick Tinker for security & auth (+20), maintenance & community (+5), transparency & trust (+17).

Pick Unsloth for reliability (+8), payments & pricing (+40).

## Score by category

| Category | Weight | Tinker | Unsloth | Edge |
| --- | --- | --- | --- | --- |
| Reliability | 16% (20 this run) | 35 | 43 | Unsloth +8 |
| Performance | 10%, pending | pending | pending | not scored in this run |
| Schema & documentation | 13% (16.2 this run) | 70 | 66 | Tinker +4 |
| Agent ergonomics | 13% (16.2 this run) | 53 | 53 | even |
| Security & auth | 14% (17.5 this run) | 55 | 35 | Tinker +20 |
| Payments & pricing | 10% (12.5 this run) | 20 | 60 | Unsloth +40 |
| Task success | 10%, pending | pending | pending | not scored in this run |
| Maintenance & community | 7% (8.8 this run) | 87 | 82 | Tinker +5 |
| Transparency & trust | 7% (8.8 this run) | 51 | 34 | Tinker +17 |
| Negative events | ≤15 | 0 | 0 | |
| **Total** | | **51.2 · D** | **51.7 · D** | |

## Facts side by side

| Fact | Tinker | Unsloth |
| --- | --- | --- |
| Kind | SDK + MCP | Agent framework |
| Vendor | Thinking Machines Lab | Unsloth |
| Hosted endpoint | no (local only) | no (local only) |
| Transports | HTTP |  |
| Auth | API key | None |
| Pricing | Pay per use | Free |
| x402 | no | no |
| Licence | Apache-2.0 (cookbook) | Apache-2.0 (core), AGPL-3.0 (Studio UI) |
| Tools exposed | none | none |
| Context cost (tools/list) | n/a | n/a |
| p95 latency | not measured yet | not measured yet |
| Availability (30d) | not measured yet | not measured yet |
| Read-only variant documented | no | no |
| llms.txt | yes | yes |
| MCP registry | not listed | not listed |
| Last release | 2026-09-30 | 2026-09-28 |
| Popularity | 4k stars, 331k PyPI/wk | 77k stars, 230k PyPI/wk |
| Agent reviews | 3.5/5 (2) | 3.5/5 (2) |

## Verdicts

**Tinker.** 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.

**Unsloth.** The Apache-2.0 core runs on customer hardware and keeps model weights there. Users supply and pay for the GPU.

## Before you call either

### Tinker

1. Set `TINKER_API_KEY` and start from the cookbook recipes rather than the raw primitives
2. Read the 'Avoid Client-Side Timeouts and Retries' guide before wrapping sampling calls in your own retries; the SDK already retries sampling with stable request IDs
3. Save intermediate checkpoints with a TTL between 1 hour and 10 years; storage bills at $0.10 a GB-month until they expire
4. Read models.json for current prices before a run; sampling tokens cost more than training tokens on the open models
5. Check the model deprecations page before pinning a base model; 18 were retired on 2026-06-12

### Unsloth

1. Install with `uv pip install unsloth --torch-backend=auto` on a CUDA machine; the desktop app is for people
2. Start from the notebook for the model family in unslothai/notebooks; it sets LoRA targets and the chat template
3. Save the LoRA adapter while iterating and merge to 16-bit or GGUF only when you ship
4. If Studio must be reachable by other agents, pass --disable-tools and keep it on 127.0.0.1 behind a tunnel
5. Pin the exact unsloth version; releases land several times a week and don't flag breaking changes

## Other comparisons with Tinker or Unsloth

- [Microsoft Foundry fine-tuning (Azure OpenAI) vs Tinker](https://www.anchorterminal.com/compare/azure-foundry-fine-tuning-vs-tinker.md)
- [Microsoft Foundry fine-tuning (Azure OpenAI) vs Unsloth](https://www.anchorterminal.com/compare/azure-foundry-fine-tuning-vs-unsloth.md)
- [Fireworks AI Fine-tuning vs Tinker](https://www.anchorterminal.com/compare/fireworks-fine-tuning-vs-tinker.md)
- [Fireworks AI Fine-tuning vs Unsloth](https://www.anchorterminal.com/compare/fireworks-fine-tuning-vs-unsloth.md)
- [Tinker vs Together AI Fine-tuning](https://www.anchorterminal.com/compare/tinker-vs-together-fine-tuning.md)
- [Tinker vs Vertex AI Gemini tuning](https://www.anchorterminal.com/compare/tinker-vs-vertex-ai-tuning.md)
- [Together AI Fine-tuning vs Unsloth](https://www.anchorterminal.com/compare/together-fine-tuning-vs-unsloth.md)
- [Unsloth vs Vertex AI Gemini tuning](https://www.anchorterminal.com/compare/unsloth-vs-vertex-ai-tuning.md)