---
title: "takuaka (@takuaka)"
description: "Rank #652 of 1133 on the Superset leaderboard, tier 1."
canonical: https://superset.sh/takuaka
last-updated: 2026-09-07
---

# takuaka (@takuaka)

Rank #652 of 1133 on the Superset leaderboard, tier 1.

URL: https://superset.sh/takuaka

## Standing

- Rank: #652 of 1133
- Tier: Button pusher
- Tokens: 3.84B all time
- Cost: $4265.55 API-equivalent
- Sessions: 576
- Tracking: 2026-07-29 to 2026-09-07

## How this tier was reached

Tier is a weighted score across the axes measured for you, so a strong axis offsets a weak one. Operator needs a higher combined score, and cost also caps the tier on its own.
Scored on 5 axes, measured over the trailing 30 days.

| Axis | Current | Needed | Met | Measures |
| --- | --- | --- | --- | --- |
| width | 1.0 | 3.1 | no | agents running at once |
| depth | 5.6M | 10.4M | no | tokens per session |
| output | 11.5 | 2.1 | yes | agent PRs merged per week |
| sustain | 30 | 17 | yes | active days in the last 30 |
| cost | $82.37 | <= $315.60 | yes | USD per merged PR, lower is better |

## Achievements

- **efficient** (level 3 of 3) at $100, earned 2026-09-07
- **ship-it** (level 2 of 4) at 10, earned 2026-09-07
- **thirty** (level 1 of 3) at 30, earned 2026-09-07
- **two-hands** (level 1 of 3) at 10, earned 2026-09-07
- **whole-task** (level 1 of 3) at 2.5M, earned 2026-09-07

## Milestones

- sessions: 500
- spend: 1.0K
- tokens: 1.00B

## Contributions

- Active days in the last year: 40
- Busiest day: 2026-09-07 at 261.6M tokens

## Models

| Provider | Model | Tokens | Cost |
| --- | --- | --- | --- |
| claude | claude-opus-4-8 | 2.33B | $2289.68 |
| claude | claude-fable-5 | 860.9M | $1426.10 |
| claude | claude-opus-5 | 541.9M | $456.46 |
| claude | claude-haiku-4-5-20251001 | 61.9M | $15.87 |
| claude | claude-opus-4-7 | 27.7M | $70.49 |
| codex | gpt-5.4 | 14.3M | $6.74 |
| codex | gpt-6-astra | 4.0M | $0.21 |

## Token breakdown

- Uncached input: 4.9M
- Cached input: 3.68B
- Cache write 5m: 11.0M
- Cache write 1h: 131.1M
- Output: 15.8M
- Reasoning output: 2.1M

Published voluntarily by the account holder. Source: https://superset.sh/takuaka
