Tool Use & Agents · sub-raceClear lead
Tool Use & Function Calling race
Choosing the right tool, filling arguments correctly and following policies across multi-turn tasks.
- Leader
- Opus 5.5
- Margin
- 1.5 pts
- 1.6% ahead
- Weeks at #1
- 4
- in a row
- Changes
- 8
- in 2026
Warming up the track…
Oct 19, 2026Opus 5.5 leads
2025Apr ’25Jul ’25Oct ’252026Apr ’26Jul ’26Oct ’26
Score history
Top 6 today over the leader's frontier · click to jump to a week
#1 frontierOpus 5.5Sonnet 5.5GPT-6.1 SolFable 5.1GPT-6 AstraGemini 4 Argon
Reigns at #1
Longest: o3, 14 weeks (from Apr 2025).
Sibling races
More tool use & agents races
Why this race matters
Every agent and integration depends on reliable function calling.
How it is scored
Pass^k success rate across simulated customer tasks.
Unit: percentHigher is betterPhoto finish under 0.5 ptsDominant over 3.0 pts
Data provenance
Retail, airline and telecom agent tasks.
MCP Atlas
Tool use across real MCP servers.
Scores here are illustrative placeholders that mimic real benchmark patterns. Swap in live data through the score provider. Methodology