Updated 2026-06-20
Which AI is best
for Scandinavian?
We evaluate over 350 AI models every morning and present the best. Mistral: Mistral Small 3 currently leads with 8.9/10.
What do you need to do?
Best choice
Meta: Llama 3.2 3B Instruct
meta-llama
$0.05/1M tokens
Cheapest option
Meta: Llama 3.2 3B Instruct
meta-llama
$0.05/1M tokens
Free alternative
Nearest free alternative
Cohere: Command R+ (08-2024)
Free (with limits)
Which AI are you using today?
AI orchestrator: Different models excel at different tasks. Use our results to choose the right model for writing, coding, research and API cost. See orchestrator ranking →
Today's ranking
See full benchmark → Tested 2026-06-20
| # | Model | Tier | Nordic | Instr | Score |
|---|---|---|---|---|---|
| 1 | Mistral: Mistral Small 3 mistralai | Budget | 8.0 | 10.0 | 8.9 |
| 2 | Google: Gemma 3 4B google | Budget | 8.0 | 10.0 | 8.7 |
| 3 | Mistral Large 2407 mistralai | Mid-range | 8.0 | 10.0 | 8.5 |
| 4 | Cohere: Command R+ (08-2024) cohere | Premium | 8.0 | 10.0 | 8.4 |
| 5 | OpenAI: GPT-5.4 Nano openai | Mid-range | 8.0 | 10.0 | 8.2 |
| 6 | Meta: Llama 3.2 3B Instruct meta-llama | Budget | 8.0 | 8.0 | 8.2 |
| 7 | Cohere: Command R7B (12-2024) cohere | Budget | 8.0 | 10.0 | 7.9 |
| 8 | Anthropic: Claude Sonnet 4.6 anthropic | Premium | 8.0 | 10.0 | 7.4 |
| 9 | Perplexity: Sonar Pro Search perplexity | Premium | 8.0 | 10.0 | 7.2 |
| 10 | OpenAI: GPT-5.4 Image 2 openai | Premium | 6.0 | 10.0 | 7.2 |
| 11 | Anthropic: Claude Haiku 4.5 anthropic | Mid-range | 6.0 | 10.0 | 6.7 |
| 12 | Google: Gemini 3.1 Pro Preview google | Mid-range | 8.0 | 4.0 | 5.3 |
Trend over last 7 days Full history →
Mistral Small 3 Gemma 3 4B Mistral Large 2407 Command R+ GPT-5.4 Nano
The chart shows direction — exact figures are for paying customers only
Latest events
20. JunNew daily leader — Mistral: Mistral Small 3 with score 8.9
20. JunNew model: OpenAI: GPT-5.4 Nano debuts with score 8.2
20. JunNew model: Mistral: Mistral Small 3 debuts with score 8.9
20. JunNew model: Meta: Llama 3.2 3B Instruct debuts with score 8.2
20. JunGemini 3.1 Pro Preview improved from 2.6 to 5.3