Apertus v1.5 8B

# Fix list: Apertus v1.5 8B

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/apertus-v1-5-8b, the October 2026 research run, assessed 21 October 2026. Grade C, 61 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Apertus v1.5 8B: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Reach, 40 out of 100, up to 23.1 more on the total

Why it scored 40: Memory 25 of 30, the BF16 files are 18.4 GB, or 22.1 GB with a fifth added, which fits 24 GB. Runtimes 0, no released vLLM, SGLang, Transformers, llama.cpp, Ollama or LM Studio runs version 1.5 yet, only Swiss AI's forks. Builds 0, Swiss AI publishes no quantised build. Terms 10 of 15, Apache 2.0 with an acceptable use policy added at the gate. Download 5, a click-through that needs an account and your contact details. Hosted 0, no inference provider or publisher API was found, and the site that lists cloud providers couldn't be reached.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reach):

Whether an agent can run it, and where.

- 0 to 30, memory. The smallest build from the publisher or a runtime's own library that keeps 4-bit precision or better, at its file size plus a fifth for the runtime and a working context. 30 for 16 GB or less, 25 for 24 GB, 20 for 32 GB, 15 for 64 GB, 10 for what fits our reference machine (the GPU share of 128 GB of unified memory), 5 for 256 GB, 0 above.
- 0 to 25, runtimes. 5 for each of llama.cpp, Ollama, vLLM, SGLang, LM Studio and Hugging Face Transformers that runs the architecture in a released version, from the runtime's own code, library or documentation. A publisher's fork or an unreleased branch doesn't count. 25 at most.
- 0 to 15, builds from the publisher. 15 for quantised builds at two precisions or more (FP8, 4-bit, GGUF), 10 for one, 0 when only other people's quantisations exist.
- 0 to 15, terms of use. 15 for an OSI-approved licence with nothing added, 10 for a custom licence that lets anyone use it commercially for free below a threshold or under a use policy, 5 for research or non-commercial use only, 0 for no licence or unclear terms.
- 0 to 10, download. 10 when the files download with no account and no terms to accept, 5 for a click-through that needs an account, 0 when a person approves each request.
- 5, hosted too. The publisher's own API serves it, or at least two inference providers do, so an agent without the hardware can still call it.

When our reference-machine runs start, speed on that machine (tokens per second and time to first token at the build that fits) joins Reach.

## 2. Risk, 75 out of 100, up to 7.7 more on the total

Why it scored 75: File safety 25, safetensors only and no code in the repository. Provenance 15, the files are in the swiss-ai organisation the report links, with a checksum on every file. Safety evaluation 10 of 20, published safety benchmarks run with Swiss AI's open harnesses and early access for three outside evaluators, but the manual evaluation is Swiss AI's own and can't be rerun. Agent guidance 10, the report says what the refusals cover, how far jailbreaks get and that capability limits are the main defence, with a personal-data filter to apply, but nothing on prompt injection for deployers. Licence risk 10, 5 off for the use policy. Security programme 5, ETH Zurich's disclosure policy, with no advisories found.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-risk):

What could go wrong when you run it, and what the publisher has done about it. A higher score means less risk.

- 0 to 25, file safety. 25 for weights in safetensors or GGUF only, 15 when pickle files (.bin, .pt, .pth) sit beside them, 0 for pickle only. 10 off when the documented way to load it or to build its prompts runs code from the repository (`trust_remote_code`, or the publisher's scripts in place of a template).
- 0 to 15, provenance. 10 when the files come from the publisher's own organisation, the one its site or docs link to, and 5 for a checksum on every file.
- 0 to 20, safety evaluation. 20 when results are published and someone other than the publisher ran them or can rerun them, 10 for the publisher's own results, 0 for none.
- 0 to 15, guidance for agent use. 15 for documented prompt-injection guidance or an instruction hierarchy the model was trained to follow, 10 for a statement of what its trained refusals cover and the safeguards to add around it, 5 for general cautions, 0 for none.
- 0 to 15, licence risk. 15 to start, 5 off for each of these, a use policy or field-of-use limits beyond the law, terms the licensor can change or revoke, a revenue or user threshold, and a licence that changed without a dated note or whose link is broken. Not below 0.
- 0 to 10, a security programme. 5 for a way to report a flaw (security.txt, a disclosure policy or a security contact), 5 for advisories handled in public.

When our reference-machine runs start, measured robustness (prompt injection with AgentDojo, over-refusal with XSTest) joins Risk.

## 3. Documentation and upkeep, 84 out of 100, up to 2.5 more on the total

Why it scored 84: Model card 14 of 20, languages and a knowledge cut-off are missing. Training data 20, every dataset is public and listed with versions, licences and counts. Last change 20, the repository changed 7 days ago. Responsiveness 20, Swiss AI answered all three questions. Version policy 5, earlier models stay in their own repositories without a dated note. File health 5, the gated config couldn't be checked against the card, and the card and the policy disagree on the output filter.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-upkeep):

- 0 to 20, the model card. Intended use, limitations, architecture, parameters, context length, languages and knowledge cut-off. 20 for all seven, 3 off for each that's missing.
- 0 to 20, training data. 20 for the data published or listed with sources and sizes, 15 for a data card that names the sources or the mix, the cut-off and the filtering, 10 for a general description with a cut-off, 5 without one, 0 for none.
- 0 to 20, the last change to the files, the template or the reference code. 20 within 30 days, 15 within 90, 10 within 180, 5 within a year, 0 older.
- 0 to 20, responsiveness. The publisher's replies to the 40 most recent questions and reports in the repository's discussions or issues. 20 for most, 10 for some, 5 for a few, 0 for none.
- 0 to 10, version policy. 10 for earlier versions kept in their own repositories with a dated note when one is replaced, 5 for earlier versions kept without notes or a first release, 0 when files were replaced in place without a note.
- 0 to 10, file health. 10 when config, tokenizer, template and card agree and no report of a broken file waits without a reply, 5 when one does, 0 when the files contradict the card.

## 4. Agent interface, 85 out of 100, up to 2.3 more on the total

Why it scored 85: Chat template 20, a Jinja file in the repository. Tool-calling format 25, Swiss AI's report documents how tools are declared and returned and the card names vLLM's apertus parser. Trained tool calling 20, trained on verified tool use with several calls a turn, though parallel calls are weaker on the 8B and tools don't work with thinking on. Structured output 0, not documented. Reasoning control 10, thinking turns on per request and the card says where the reasoning comes back. Context 10, 262,144 tokens trained.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-interface):

What a model and a developer need to talk to it. Speech, image and other models with no chat interface read the tool-calling and structured-output lines for what the model takes and returns (voices, timestamps, streaming, typed outputs), and the note says how.

- 0 to 20, the chat template. 20 for a template in the repository (Jinja, in the tokenizer config or its own file), 10 when the prompt format is only given as code you run.
- 0 to 25, the tool-calling format. 25 when the publisher documents how tools are declared and how calls come back, with the parser to use in a named runtime, 10 when only the template or a runtime's own recipe says it.
- 0 to 20, tool calling the model was trained for. 20 with several calls in one turn, 10 for one call at a time.
- 0 to 15, structured output. 15 when JSON or JSON Schema output is documented for the model with how to ask for it, 5 when the card claims it without saying how.
- 10, reasoning control. Thinking that can be turned off or budgeted per request, with where the reasoning text comes back.
- 0 to 10, context. 10 for 128K tokens or more trained, 5 for 32K to 127K. Context reached only by scaling the position encoding counts one band down.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- unchecked, apertus-ai.org (documentation, providers and contact pages) didn't answer through our proxy, so hosted providers and any security contact there are scored as absent
- unchecked, config.json, the tokenizer config and the commit history are behind the gate
- Runtimes will move as soon as vLLM or Transformers ships Apertus 1.5 support, which the card says is in progress
- Safety evaluation sits between 10 and 20, since most results can be rerun with Swiss AI's harnesses and outside bodies had early access, but the manual evaluation can't be rerun
- Security programme credits ETH Zurich's disclosure policy, which is scoped to ETH's IT infrastructure rather than the model

## Weaknesses

- No released runtime supports it, only Swiss AI's forks of vLLM and Transformers
- No quantised build from Swiss AI, and the report warns 4-bit NVFP4 in attention gives NaN logits
- Gated download that collects contact details, under a use policy you must pass on when you redistribute
- Tool calling and thinking don't combine, and parallel calls are weaker on the 8B
- No structured output documentation and no knowledge cut-off

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Accept the terms on the card and set HF_TOKEN before downloading
- Use Swiss AI's Docker image ghcr.io/swiss-ai/vllm_apertus_1.5_release, since stock vLLM and Transformers don't load Apertus1p5ForConditionalGeneration yet
- Serve with `--chat-template-content-format string --enable-auto-tool-choice --tool-call-parser apertus` for tools, and leave thinking off when you need tools
- For thinking, set `--reasoning-parser apertus` and `--default-chat-template-kwargs.enable_thinking true`, and budget max tokens generously
- Audio is resampled to 24 kHz and costs 40 tokens a second, and video must be passed as frames
- Don't cast the loaded model with .half(), the image and audio tokenizers must stay in F32
- If you redistribute the weights, the policy says you must pass it on and get the recipient's express assent

## What the review panel asked for

- date the supersession
- publish config outside the gate
- upstream runtime support
- a 4-bit build
- tool calling with thinking on
- a model-specific security contact
- prompt-injection guidance for deployers

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.