Sign in
  1. 1Claude Opus 5.5The top Artificial Analysis score of these 5.$4 / $20
  2. 2Claude Fable 5.1Also #2 for writing.$10 / $50
  3. 2Gemini 4 ArgonAlso #2 for coding.$2 / $10
  4. 4Claude Sonnet 5.5The fastest of these 5, at 138 tokens a second.$2 / $10
  5. 5GLM 5.3 FlashThe lowest API price of these 5.$0.15 / $0.50

Top 10 for research

Scores from artificialanalysis.ai and lmarena.ai.

API price is per 1M tokens, in / out. Updated Nov 2, 2026.

The 10 models that rank best for research
#ModelArtificial AnalysisLMArena textAPI price
1Claude Opus 5.557.61,516$4 / $20
2Claude Fable 5.153.41,513$10 / $50
2Gemini 4 Argon——$2 / $10
4Claude Sonnet 5.556.01,473$2 / $10
5GLM 5.3 Flash41.81,472$0.15 / $0.50
6Kimi K343.61,475$3 / $15
7GLM 5.344.81,474$1.40 / $4.40
8GPT-6 Astra52.7—$10 / $50
9Muse Spark 1.348.11,491$1.25 / $4.25
10MiMo V2.6 Pro46.31,491$0.43 / $0.87

We rank every model per task, blending public leaderboards with our own blind tests where we have them. Price and speed don’t count toward the rank. How we rank

Keep exploring

Ready when you are.

Try Claude Opus 5.5