Models · compare

Compare the models, side by side.

Pick two or three of Evanesca's models to see their facts next to each other: where they run, the checks they passed, context, tool calling and price. Further down, one table holds all 17 models, every one run in Europe — in France, in Mistral AI's EU region, in Germany and in Switzerland.

Checked on 20 Oct 2026

Side by side

The best value on each row is marked when the models differ. Each column's name leads to the model's page.

First modelSecond modelThird model

COMPARISON · PRICES VAT INCLUDED
Fact	DeepSeek V4 Flash Paris, France	Qwen3.5 397B Paris, France	Mistral Small 3.2 Paris, France
Maker	DeepSeek	Qwen	Mistral AI
Run by	Scaleway	Scaleway	Scaleway
Checks	6/6	6/6	5/6 · Security ✕
Licence	MIT	Apache-2.0	Apache-2.0
Open weights	Yes	Yes	Yes
Size	284B (13B active)	397B (17B active)	24B
Context	256k tokens	250k tokens	Not verified
Type	Text	Text + images	Text + images
Tool calling (measured)	Yes	Yes	Yes
Reasoning	Yes	Yes	—
Input / M tokens	€0.672	€1.008	€0.252 lowest
Output / M tokens	€1.344	€6.048	€0.588 lowest
Quick question (300 tokens in, 400 out)	€0.0007	€0.0027	€0.0003 lowest
Model card	huggingface.co	huggingface.co	huggingface.co

Every model in one table

All 17 models, every one run in Europe. Prices per million tokens, VAT included; “Tools”: tool calling measured by a real call.

EVERY MODEL · COMPARISON
Model	Where	Checks	Context	Tools	Type	Input / M	Output / M	Quick question
Mistral Large 4	EU region of Mistral AI	4/6 · Licence ✕ · Provenance ✕	512k	Yes	Text + images	€2.5133	€7.7246	€0.0038
DeepSeek V4.1 Flash	Germany (AKI.IO)	6/6	1M	Yes	Text + images	€1.008	€3.024	€0.0015
GLM-5.3	Germany (AKI.IO)	6/6	512k	Yes	Text	€3.024	€9.24	€0.0046
Qwen3.8 27B	Paris, France	6/6	256k	Yes	Text + images	€1.008	€5.544	€0.0025
DeepSeek V4 Flash	Paris, France	6/6	256k	Yes	Text	€0.672	€1.344	€0.0007
Apertus v1.5 70B	Switzerland (Infomaniak)	4/6 · Security, Benchmarks: not run yet	100k	Yes	Text + images	€1.2936	€4.62	€0.0022
GLM-5.2	Paris, France	6/6	256k	Yes	Text	€3.024	€9.24	€0.0046
Qwen3.6 35B	Paris, France	6/6	256k	Yes	Text + images	€0.42	€2.52	€0.0011
Gemma 4 26B	Germany (AKI.IO)	6/6	256k	Yes	Text + images	€0.168	€0.84	€0.0004
Mistral Small 4	EU region of Mistral AI	5/6 · Security ✕	256k	Yes	Text + images	€0.2772	€1.1088	€0.0005
Qwen3.5 397B	Paris, France	6/6	250k	Yes	Text + images	€1.008	€6.048	€0.0027
GPT-OSS 120B	Paris, France	5/6 · Security ✕	—	Yes	Text	€0.252	€1.008	€0.0005
Qwen3 Coder 30B	Paris, France	5/6 · Security ✕	256k	Yes	Code	€0.336	€1.344	€0.0006
Qwen3 235B	Paris, France	5/6 · Security ✕	250k	Yes	Text	€1.26	€3.78	€0.0019
Mistral Small 3.2	Paris, France	5/6 · Security ✕	—	Yes	Text + images	€0.252	€0.588	€0.0003
Llama 3.3 70B	Paris, France	5/6 · Security ✕	—	Yes	Text	€1.512	€1.512	€0.0011
Pixtral 12B	Paris, France	5/6 · Security ✕	128k	Yes	Vision	€0.336	€0.336	€0.0002

Quick question: 300 tokens in and 400 out. A missing context is not verified yet. Reasoning tokens count as output. “not run yet”: checks not run on that model so far — not failed ones. How billing works · How the checks are run

Quick answers

Lowest cost

Pixtral 12B (Paris, France): €0.0002 for a quick question (300 tokens in, 400 out).

Reads images

Mistral Large 4, DeepSeek V4.1 Flash, Qwen3.8 27B, Apertus v1.5 70B, Qwen3.6 35B, Gemma 4 26B, Mistral Small 4, Qwen3.5 397B, Mistral Small 3.2 and Pixtral 12B.

Longest context

1M tokens, as the operator announces it: DeepSeek V4.1 Flash.

All six checks

DeepSeek V4.1 Flash, GLM-5.3, Qwen3.8 27B, DeepSeek V4 Flash, GLM-5.2, Qwen3.6 35B, Gemma 4 26B and Qwen3.5 397B.