Tool Use & Agents · compositePhoto finish
Tool Use & Agents race
Calling tools reliably, operating computers, researching the web and staying on task for hours.
- Leader
- Opus 5.5
- Margin
- 0.3 pts
- 0.3% ahead
- Weeks at #1
- 4
- in a row
- Changes
- 7
- in 2026
Warming up the track…
Oct 19, 2026Opus 5.5 leads
2025Apr ’25Jul ’25Oct ’252026Apr ’26Jul ’26Oct ’26
Re-weight this race
Drag to say what matters to you. The track re-sorts live and weeks-at-#1 are recomputed for your weighting.
23%
23%
23%
30%
Score history
Top 6 today over the leader's frontier · click to jump to a week
#1 frontierOpus 5.5GPT-6 AstraFable 5.1Sonnet 5.5GPT-6.1 SolGemini 4 Argon
Reigns at #1
Longest: o3, 14 weeks (from Apr 2025).
Inside this race
What makes up tool use & agents
Why this race matters
Agents turn models from advisers into workers. This race decides who powers them.
How it is scored
Normalised subcategory scores, weighted toward long-horizon autonomy.
Unit: indexHigher is betterPhoto finish under 0.6 ptsDominant over 3.5 pts
Data provenance
Tool-using customer-service agents.
Real desktop tasks.
Length of tasks agents complete.
Scores here are illustrative placeholders that mimic real benchmark patterns. Swap in live data through the score provider. Methodology