LLM Reference

The research leaderboard · for knowledge workers

Best for research

5 editor picks · 9 eligible models · Deep reading, citations, synthesis.

Editorial pick plus benchmark and API pricing context.

See raw /best
EDITOR'S CHOICEResearched 7d ago

Claude Fable 5

Anthropic · 1m context
Excellent

Deep reading and footnoted synthesis the experts respect.

GDPval-AA ELO 1932 and Anthropic-reported finance, trading, and analytics wins make it the strongest general knowledge-work pick; do not use Mythos-only HLE rows as Fable evidence.

Open model
The numbers
$/1M out
$50.00
$10.00 input
Context
1m
max window
Pros
  • +GDPval-AA ELO 1932
  • +Strong long-horizon document analysis
  • +1M context
Cons
  • −$50 / 1M out
  • −No Fable-specific GPQA or HLE score yet

Also worth picking

The runners-up

ranked by editorial pick orderEditorial tiersExcellentStrongSolid
#2
Claude Opus 4.7
Anthropic · 1m
$25.00 / 1M out
Still strong for writing-adjacent research and everyday synthesis at lower cost.
#3
GPT-5.5
OpenAI · 1.05m
$30.00 / 1M out
GPQA 93.6 with a high reasoning ceiling; strongest on quantitative literature.
#4
Gemini 3 Pro
Google DeepMind · 1m
$5.00 / 1M out
GPQA 91.9 plus live search and a 1M window — best for tracking the literature in real time.
#5
Grok 4.3
xAI · 1m
$2.50 / 1M out
GPQA 90.1, fresh, 1M context with real-time retrieval baked in.

Eligibility

9 models are eligible for this board

Eligibility means tagged with useCases: [research]. Pins must come from this pool.

All picks