🏆 Workload Leaderboard · 2,000 tasks/month benchmark
Best LLM Models for 200k–10M Long-Context Repository & Document Sweeps (2026 True-Cost Ranking)
Whole-monorepo security scans and M&A due diligence where crossing 200k tokens triggers 2x long-context cliff penalties on many frontier APIs.
3 van 4 Featured Sponsor Slots Vrij
Bouw jij een tool voor Long-Context (>200k)?
Claim een sticky Featured Listing bovenaan dit leaderboard + op gerelateerde modelpagina's (cap: 50.000 impressies/mnd, €99/mnd via Stripe via Grok Bot).
| Rank | Model | Monthly True Cost | Unit Cost / 1k Tasks | Code / Reason | Latency · Speed | Action |
|---|---|---|---|---|---|---|
| #1 | Llama 4 Scout 109B (Meta (Hosted Cloud)) | €74 ($80) | $40.24 / 1k tasks | 83.4 / 82.1 | 170ms · 260 t/s | Deploy → |
| #2 | Gemini 4 Argon (Google DeepMind) | €241 ($262) | $130.94 / 1k tasks | 89.4 / 88.1 | 210ms · 195 t/s | Deploy → |
| #3 | Gemini 4 Carbon (Google DeepMind) | €1,942 ($2,111) | $1055.64 / 1k tasks | 96.2 / 97.4 | 540ms · 92 t/s | Deploy → |
| #4 | Claude Sonnet 5.5 (Anthropic) | €3,372 ($3,665) | $1832.69 / 1k tasks | 95.4 / 93.8 | 430ms · 98 t/s | Deploy → |