# Tess 4 27B

URL: https://interfaze.ai/models/migtisseratess-4-27b

[All models](https://interfaze.ai/models)

Tess 4 27B by migtissera, a image-text-to-text model with multimodal capabilities. Understand and compare multimodal features, benchmarks, and capabilities.

## Comparison

| Feature | Tess 4 27B | Interfaze |
| --- | --- | --- |
| Input Modalities | text, image | image, text, audio, video, document |
| Native OCR | No | Yes |
| Long Document Processing | No | Yes |
| Language Support | unknown | 162+ |
| Native Speech-to-Text | No | Yes |
| Native Object Detection | No | Yes |
| Guardrail Controls | No | Yes |
| Context Input Size | 64K | 1M |
| Tool Calling | Yes | Tool calling supported + built in browser, code execution and web search |

### Scaling

| Feature | Tess 4 27B | Interfaze |
| --- | --- | --- |
| Scaling | Self-hosted/Provider-hosted with quantization | Unlimited |

[Try Interfaze](https://interfaze.ai/dashboard)[Read the Docs](https://interfaze.ai/docs)

View model card on [Hugging Face](https://huggingface.co/migtissera/Tess-4-27B)

> **Reasoning that scales with the problem.** An agentic, thinking-native model that deliberates _harder exactly when it matters_ — and gets out of its own way when it doesn't.

**Tess-4-27B** is the first Tess release in two years, and the first that _reasons_. Built on **[Qwen/Qwen3.6-27B](https://huggingface.co/Qwen/Qwen3.6-27B)** by **[Migel Tissera](https://huggingface.co/migtissera)**, it's post-trained on a deliberate blend: 64K-token long-context agentic traces — real engineering work done with **Fable-5**, not synthetic generations — with a reasoning style approximated from Fable-5 by a three-model teacher ensemble (**Opus-4.8**, **GPT-5.5**, and **GLM-5.2**) fused into one coherent voice.

The result is a 27B model that thinks like a senior engineer: form a hypothesis, act, verify, and reason with real density on the turns that actually deserve it — **not a model that narrates its way to an answer it already had.**

* * *

## Community Performed Benchmarks

**Currently best-in-class for BenchLocal**

| Rank | Model | Score | Result |
| --- | --- | --- | --- |
| 1 | Tess-4-27B (Q8) | 81% | 122/150 |
| 2 | Qwen3.6-35B-A3B (UD-Q8\_K\_XL) | 78% | 117/150 |
| 3 | Gemma-4-31B (Q6 · 180k ctx) | 78% | 117/150 |
| 4 | Qwopus3.6-27B Coder-Compat (Q6\_K) | 77% | 116/150 |
| 5 | Qwen3.6-27B pi-tune (Q8) | 77% | 115/150 |

### References

1.  [https://huggingface.co/migtissera/Tess-4-27B/discussions/2#6a4ff70af13ec7012fb149f0](https://huggingface.co/migtissera/Tess-4-27B/discussions/2#6a4ff70af13ec7012fb149f0)
2.  [https://gist.github.com/everson/261fdef8a3d35298b36a07f436e407f6](https://gist.github.com/everson/261fdef8a3d35298b36a07f436e407f6)

## Why Tess-4 is different

-   🧠 **Weight-scaled reasoning.** Tess-4 keeps routine steps tight and pours deliberation into the hard ones — planning, debugging, synthesis, judgment calls. It doesn't ramble; it thinks _proportionally_ to the difficulty of the moment.
-   🛠️ **Agentic by design.** Native, parallel tool use and disciplined multi-step problem solving. It reads a codebase, builds a real mental model, and acts on it.
-   📏 **Long-context, trained at 64K.** Post-trained on **64K-token long-context agentic traces**, so it holds a large working set without losing the thread.
-   👁️ **Multimodal.** Inherits Qwen3.6's vision tower — text **and** image in. (For GGUF, pair with the included vision projector.)
-   🤝 **Honest, not sycophantic.** Trained to give grounded, evidence-based pushback instead of flattery.

## The reasoning traces

Tess-4's signature is _how it thinks_. The reasoning/thinking traces used to train it were a **best-case approximation of Fable-5**, produced by a combination of **Opus-4.8, GPT-5.5, and GLM-5.2** working together as a team — a multi-model teacher ensemble distilled into a single, coherent reasoning style.

The result is a model that reasons **prospectively** — predicting, verifying, and weighing alternatives _before_ acting — rather than narrating after the fact.

## Prompt format & thinking

Tess-4 uses the Qwen3.5-family chat template with explicit `<think> … </think>` reasoning blocks. The model reasons privately, then produces its visible answer:

`<|im_start|>user Your prompt here<|im_end|> <|im_start|>assistant <think> … the model's private reasoning … </think> … the model's answer …<|im_end|>`

Apply it automatically via `tokenizer.apply_chat_template(messages, add_generation_prompt=True)`, or `--jinja` in llama.cpp.

## Available formats

**This repo — full-precision weights:**

| Format | ~Size | Best for |
| --- | --- | --- |
| BF16 safetensors | 52 GB | transformers · vLLM · SGLang |

**GGUF quants → [`migtissera/Tess-4-27B-GGUF`](https://huggingface.co/migtissera/Tess-4-27B-GGUF)**

| File | Format | ~Size | Best for |
| --- | --- | --- | --- |
| Tess-4-27B-Q4\_K\_M.gguf | Q4\_K\_M | 16.5 GB | smallest — great quality/size · most popular |
| Tess-4-27B-Q6\_K.gguf | Q6\_K | 22 GB | near-lossless |
| Tess-4-27B-Q8\_0.gguf | Q8\_0 | 28 GB | effectively lossless |
| mmproj-Tess-4-27B-F16.gguf | vision projector | 0.9 GB | pair with any text GGUF for image input |

## Faster inference

-   ⚡ **[Tess-4-27B-EAGLE3](https://huggingface.co/migtissera/Tess-4-27B-EAGLE3)** — a speculative-decoding draft trained on Tess-4's own outputs: **1.76× average decode speedup, up to 2.4× on reasoning** (measured on H100; lossless — outputs are identical). SGLang: `--speculative-algorithm EAGLE3 --speculative-draft-model-path migtissera/Tess-4-27B-EAGLE3`; vLLM: `--speculative-config '{"method":"eagle3","model":"migtissera/Tess-4-27B-EAGLE3","num_speculative_tokens":4}'`.
-   🧮 **[Tess-4-27B-NVFP4](https://huggingface.co/migtissera/Tess-4-27B-NVFP4)** — 4-bit NVFP4 (19 GB, −63%), Blackwell-native W4A4, calibrated on Tess-4's own generations. Quantization and speculative decoding **stack**.

## Quickstart

### llama.cpp / LM Studio (GGUF)

Grab the quant(s) from [`migtissera/Tess-4-27B-GGUF`](https://huggingface.co/migtissera/Tess-4-27B-GGUF):

```
hf download migtissera/Tess-4-27B-GGUF \
  Tess-4-27B-Q4_K_M.gguf mmproj-Tess-4-27B-F16.gguf \
  --local-dir ./tess-4-27b
```

```
llama-cli -m Tess-4-27B-Q4_K_M.gguf --jinja -p "Refactor this function and explain your reasoning."


llama-mtmd-cli -m Tess-4-27B-Q4_K_M.gguf \
  --mmproj mmproj-Tess-4-27B-F16.gguf \
  --image photo.png -p "What's in this image?"
```

**LM Studio:** put `mmproj-Tess-4-27B-F16.gguf` in the **same folder** as the model file — LM Studio auto-detects it and enables image input. (Use a recent runtime; older llama.cpp builds won't recognize the architecture.)

### transformers

```
from transformers import AutoProcessor, AutoModelForImageTextToText
import torch

model_id = "migtissera/Tess-4-27B"
processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForImageTextToText.from_pretrained(
    model_id, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True
)

messages = [{"role": "user", "content": "Explain the tradeoffs of LoRA vs full fine-tuning."}]
inputs = processor.apply_chat_template(
    messages, add_generation_prompt=True, return_tensors="pt"
).to(model.device)

out = model.generate(inputs, max_new_tokens=1024)
print(processor.decode(out[0], skip_special_tokens=True))
```

_(Requires a recent `transformers` with Qwen3.5/3.6 support.)_

## What it's good at

-   **Agentic coding** — exploring unfamiliar repos, planning changes, and executing multi-step work with tools.
-   **Long-context work** — reasoning over large codebases and documents without dropping context.
-   **Technical & product judgment** — honest, structured analysis that pushes back with evidence rather than agreeing by default.

## Credits

Tess-4-27B is built on **[Qwen/Qwen3.6-27B](https://huggingface.co/Qwen/Qwen3.6-27B)** by the **Qwen team** — full credit to them for an outstanding base model. Tess-4 inherits its Qwen3.5-family vision-language architecture and its **Apache 2.0** license.

## License

Released under the **Apache License 2.0**, inherited from the base model. See [`LICENSE`](https://huggingface.co/migtissera/Tess-4-27B/blob/main/LICENSE).

## Citation

```
@misc{tissera2026tess4,
  title        = {Tess-4-27B},
  author       = {Migel Tissera},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/migtissera/Tess-4-27B}},
  note         = {Built on Qwen/Qwen3.6-27B}
}
```

* * *

_Tess-4-27B — part of the **Tess** series by [Migel Tissera](https://huggingface.co/migtissera). Evaluations forthcoming._

## Want more deterministic results?

[Try Interfaze](https://interfaze.ai/dashboard)[Read the Docs](https://interfaze.ai/docs)