LLM Reference

The tool use leaderboard · for developers

Best for tool use

5 editor picks · 9 eligible models · Reliable function calling and structured outputs.

Editorial pick plus benchmark and API pricing context.

See raw /best
EDITOR'S CHOICEResearched 186d ago

Gemini 3 Pro

Google DeepMind · 1m context
Excellent

Structured output is table stakes — pick on schema reliability and price.

Best current BFCL (72.5) with rock-solid JSON-schema adherence and a 1M window at $5 out.

Open model
The numbers
$/1M out
$5.00
$1.25 input
Context
1m
max window
Pros
  • +Top current BFCL
  • +Native structured outputs
  • +Cheap for the tier
Cons
  • −Trails Claude on tricky tool selection

Also worth picking

The runners-up

ranked by editorial pick orderEditorial tiersExcellentStrongSolid
#2
Claude Sonnet 5
Anthropic · 1m
$10.00 / 1M out
Watchlist: strong function calling, structured outputs, 86.6% BrowseComp, 81.2% OSWorld, and 57.4% HLE-with-tools — do not claim BFCL leadership without a BFCL row.
#3
Claude Sonnet 4.6
Anthropic · 1m
$15.00 / 1M out
Best at picking the right tool when ten look plausible; pairs schema discipline with τ-bench leadership.
#4
GPT-5.4
OpenAI · 1.05m
$15.00 / 1M out
Matches GPT-5.5's structured-output reliability at half the output price ($15 vs $30) — the GPT to run at high call volume.
#5
Qwen3.5-397B-A17B
Alibaba · 262k
$2.34 / 1M out
BFCL 72.9 with open weights — strong function calling you can self-host.

Eligibility

9 models are eligible for this board

Eligibility means tagged with useCases: [tool-use]. Pins must come from this pool.

All picks