BASELINE FIXTURE — Bracket-6’s first verified stored Arena observation, accessed October 16 in America/Chicago. No rise, fall or new lead is claimed. The source header displayed October 15; observation time is recorded separately.
Claude Opus 5.5 (Max) leads the raw score table. Against GPT-6.1 Sol (Max), its advantage is 5.28 percentage points; its median task costs 8.29 times as much: $5.97 versus $0.72. Sol is the lower-cost contender in this leading-five comparison—not the cheapest model in the field. The scoreboard and the invoice award different distinctions.
THE EVENT: Agent Arena’s Overall orchestrator field. WHAT IS MEASURED: net improvement aggregates real-session signals—task success, praise/complaints, steerability, shell-tool reliability and tool hallucination. WINNING MEANS: stronger measured improvement relative to this Arena baseline, not a universal intelligence title. Median cost/task is a separate efficiency statistic, never a bonus added to the score.
Close placements remain unresolved where published intervals overlap. Astra, Fable, Sol and Argon overlap; that does not turn their raw ranks into proved differences or exact ties. Opus and Sol’s displayed intervals do not overlap. Interval overlap alone is not a formal pairwise significance test.
Raw standings sort the score. The Pareto frontier instead keeps contestants for which no observed alternative is both cheaper and at least as strong, with one strict advantage. Arena’s cost view includes Opus, Astra and Sol among several frontier models. There is no sole efficiency champion. This is a point-estimate frontier, with score uncertainty still attached.
KIT AND COST: Max and High are disclosed effort labels, not proof of equal computational budgets. The exact harness build and latency are not published in this table. Costs are P50 realized task costs over the last 14 days; they are not token list prices or a controlled same-task price test. Per-model sample counts and uncertainty remain visible below. Bracket-6 will compare later observations only within the same disclosed event, rules, category, division, scaffold and metric boundary.
BASELINE FIXTURE
First verified stored observation; no historical movement claimed.
Agent Arena · Orchestrator capabilities · Overall · Official overall mixed-effort field.
Net Improvement: Aggregated real-session performance improvement relative to Arena’s model baseline; percentage-point effects, not task-completion percentage.
Higher measured score in this event. Raw standings retain ties. No universal champion is implied.
P50 USD/task, last 14 days. Costs do not add to the score. Samples are per contestant, not the whole site total. Effort and scaffold remain part of the entrant; mixed-effort standings are not a controlled equal-budget test.
Intervals are source-reported; overlapping intervals leave close placements unresolved, not proven ties. The aggregate interval’s confidence level is unconfirmed. Latency was not published in this observation. Exact harness builds are retained when reported; unknown builds remain explicit.
. Browser header: Oct 15, 2026; accessed Oct 16 America/Chicago. Signal definitions updated 2026-10-12; methodology page updated 2026-10-15. Arena Agent Mode orchestrator; harness build not published. Official standings ↗ · Official cost / Pareto view ↗
| Raw rank | Contestant / kit | Score / uncertainty | Sessions | Cost / task |
|---|---|---|---|---|
| 1 | Claude Opus 5.5 (Max)Anthropic · Max · ProprietarySCORE LEADERPOINT-ESTIMATE FRONTIER | +18.88%±2.74 pp16.14 to 21.62% | 10,551 | $5.97USD P50 · not score |
| 2 | GPT-6 Astra (Max)OpenAI · Max · ProprietaryPOINT-ESTIMATE FRONTIER | +14.13%±2.30 pp11.83 to 16.43% | 13,168 | $3.55USD P50 · not score |
| 3 | Claude Fable 5.1 (Max)Anthropic · Max · Proprietary | +14.03%±1.99 pp12.04 to 16.02% | 16,374 | $7.55USD P50 · not score |
| 4 | GPT-6.1 Sol (Max)OpenAI · Max · ProprietaryLOWER-COST LEADING CONTENDERPOINT-ESTIMATE FRONTIER | +13.60%±1.90 pp11.70 to 15.50% | 13,444 | $0.72USD P50 · not score |
| 5 | Gemini 4 Argon (High)Google · High · Proprietary | +13.23%±1.33 pp11.90 to 14.56% | 30,489 | $0.95USD P50 · not score |
Complete observed field and recent form: Sports desk →. The report preserves the fixture observed at filing.
Public source ledger