whatmodel.sh / Apertus 1.5 70B

Apertus 1.5 70B

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Apertus 1.5 70B is an open-weights language model from Swiss AI, released July 2026. Its published size is 72.01B parameters and a 262k-token context window. Licence: apache-2.0.

Catalogue identity notes

swiss-ai/Apertus-v1.5-70B @59e744e (created 2026-07-24, apache-2.0, gated: access form). Swiss AI Apertus 1.5 70B (July 2026): continued pretraining of Apertus 1.0 with image and audio input and an optional thinking mode. Card has no score tables yet ('The technical report ... will be published in the coming weeks'). Independent row: Miðeind Icelandic leaderboard (thinking on).

Model information sources

Release date ↗ · Weights licence ↗

HF safetensors metadata 72,005,819,139 parameters (includes the image/audio encoders).

52.6±6.7#384 of 949 LLMsfrom 1 benchmark result across 1 source* thinly tested, treat the rank as provisional

$ how it compares

The LLMs ranked either side of Apertus 1.5 70B, out of 1132 with a score, then more from Swiss AI.

#modelscorewhereRAMreleased
497Hammer2.1-7b MadeAgents53±4local~5.9GB2024-12
498Gemini Experimental 1114 Google53±2comparison only2024-11
499Apertus 1.5 70B Swiss AI53±7local~153GB2026-07
500Claude 3 Opus Anthropic53±0comparison only2024-02
501LFM2.5 2.6B Liquid AI53±1local~3.2GB2026-07
more from Swiss AI
456Apertus 70B Instruct (2509) Swiss AI54±6local~22GB2025-09
650Apertus 1.5 8B Swiss AI48±7local~6.2GB2026-07
917Apertus 8B Instruct (2509) Swiss AI38±4local~6.3GB2025-08

see all LLMs · find the ones that fit your machine

$ what it's good at

other tests1 not in a task · 80% of its score
testresultvs the others in its rundifficultycountswhy
Icelandic LLM Leaderboard average (Miðeind, Sept 2026 suite)17.5% ✎▁▁▃▃▂▄▁▃▄▄▄█ beats 5/38very hard · avg 14%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 80%typical

$ where to run it

Other ways to run it

  1. Run it as published · needs ~144 GB, NVIDIA only · from huggingface.co/swiss-ai (official)
    vllm serve swiss-ai/Apertus-v1.5-70B --chat-template-content-format string --tensor-parallel-size 4 --max-model-len 262144 --reasoning-parser apertus --default-chat-template-kwargs.enable_thinking true

Official weights and documentation: huggingface.co/swiss-ai ↗

$ how the score was worked out · Miðeind Icelandic LLM Leaderb… 80%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Miðeind Icelandic LLM Leaderboard1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 80%CC BY 4.0
prior · pull toward the middle▇▇▇▇▇▇ 20%
weighted average= 52.6

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json