星海词元星海词元
Ready to build?Get an API key — all models at 50% of the official price →
Home › Capabilities › Frontier models

Frontier models

best LLM API, frontier model API, highest intelligence model

These are the strongest models in the catalogue by measured intelligence index. They cost accordingly, so they are best reserved for work where a wrong answer is expensive: complex refactors, multi-hop research, ambiguous instruction following. Route to them deliberately rather than by default.

Models (12)

ModelsProvidersInput $/MOutput $/MIntelligence indexCoding index
claude-opus-5-5 frontierAnthropic$1.00$5.0057.6—
claude-sonnet-5-5 frontierAnthropic$0.50$2.5056.0—
claude-fable-5-1Anthropic$2.50$12.5053.481.6
gpt-6-astraOpenAI$2.50$12.5052.776.9
gemini-4-argon—$2.00$10.0052.6—
gpt-6.1-solOpenAI$0.50$2.5051.8—
claude-opus-5Anthropic$1.25$6.2550.878.0
claude-fable-5Anthropic$2.50$12.5049.676.5
muse-spark-1.3-contributor frontierMeta$0.05$0.1048.175.8
gpt-6-solOpenAI$0.50$2.5047.6—
gpt-5.6-solOpenAI$1.00$5.0047.077.4
qwen3.8-maxAlibaba Qwen$0.92$2.7745.476.2

Prices are per 1M tokens in USD · ¥6.5 = $1

Capabilities