MODEL FINDER

Find the right model for the job.

Pick what you need it for and what it must do. The finder ranks every tested model on the published test for that task, within your budget.

Ranked by Arena text

People ask two anonymous models their own question and vote for the better answer. The ranking comes from millions of these blind votes.

220 of the 220 tested models meet your needs.

Best match

The highest Arena text score among the models that meet your needs.

Claude Opus 4.6Anthropic, tested as Claude Opus 4.6 (high)

Arena text: 1505#2 of 389
Price per 1M tokens: $10.00
Context: 1M

Calls tools
Reads images
Reasoning mode

Best value

Within 25 points of the best match, and 18 times cheaper.

MiMo-V2.6-ProXiaomi

Arena text: 1480#23 of 389
Price per 1M tokens: $0.544
Context: 1M

Open weights
Calls tools
Reads images
Reasoning mode

Best open weights

The best model here that you can download and run yourself.

Kimi K3Moonshot AI, tested as Kimi K3 (max)

Arena text: 1488#13 of 389
Price per 1M tokens: $6.00
Context: 1M

Open weights
Calls tools
Reads images
Reasoning mode

Also worth a look

Rank on the test: #3Claude Fable 51504$20.00 per 1M, 1M context
Rank on the test: #4Claude Opus 5.51504$8.00 per 1M, 1M context
Rank on the test: #5Claude Opus 4.71501$10.00 per 1M, 1M context
Rank on the test: #6Claude Fable 5.11501$20.00 per 1M, 1M context
Rank on the test: #7Gemini 3.8 Flash1495$1.50 per 1M, 1M context

Models are ranked by one published test at a time; scores from different tests are never combined. Prices are list prices per million tokens, three input to one output. Results published by Arena (CC BY 4.0). Every result on this test