Glean reveals model routing scores: GPT-5.6 Luna leads at $0.08
testingcatalog · x · 2026-08-27
Glean shared scoring data for its model routing system, comparing performance and cost across frontier models:
- GPT-5.6 Luna: Score 55, $0.08/task
- GLM 5.2: Score 57, $0.35
- Gemini 3.7 Flash: Score 61, $0.47
- Kimi K3: Score 63, $0.90
- Claude Opus 5: Score 67, $2.96
The system uses a specialist model called Glean Waldo (fine-tuned from NVIDIA Nemotron 3 Nano) to dynamically set reasoning levels at runtime and hand off tasks when necessary. It routes across 40+ models and supports "effort routing" to adjust thinking intensity.
More from Infra
- ComfyUI INT6 Quantization Node Cuts Storage by 25% — BakaPotatoLord · 2026-08-27
- Chip sanctions backfire? SemiAnalysis says 100T free tokens per day — basedjensen · 2026-08-27
- Foresight Institute: Open science needs open compute as private monopoly hinders independent research — niloofar_mire · 2026-08-27
- Benchmark: Minimax H3 runs on 8GB VRAM with optimized attention — Zironic · 2026-08-27
- Google reveals 9,600-chip TPU 8t; OpenAI details 3-gen chip roadmap — SumitGup · 2026-08-27
- Optimizing inference on 4090: sub-10ms latency achieved — yacineMTB · 2026-08-27