🏷️ Bulk Classification frontier
As of October 25, 2026, the highest Bulk Classification score on this frontier is Gemini 4 Argon at 100.0, with an estimated monthly cost of $1,800.00.
Classifying / extracting fields from millions of documents. Quality barely matters past a threshold; cost dominates.
The y axis is this scenario's weighted score; the x axis starts at estimated monthly cost for 50,000 requests a day. Switch axes to compare against speed, latency or context. See the full ranking.
On the frontier
| Model | Score | Est. $/month |
|---|---|---|
| Gemini 4 ArgonGoogle | 100.0 | $1,800 |
| Gemini 3.8 FlashGoogle | 92.5 | $675 |
| MiMo-V2.6-ProXiaomi (MiMo) | 90.5 | $626 |
| GLM 5.3 FlashZhipu AI (GLM) | 85.3 | $240 |
| DeepSeek V4.1 FlashDeepSeek | 83.1 | $144 |
| DeepSeek V4 Flash 0423DeepSeek | 74.2 | $131 |
| Qwen3.5-FlashAlibaba (Qwen) | 64.6 | $109 |
| GPT-6 LunaOpenAI | 62.8 | $90 |
| Gemini 2.5 Flash LiteGoogle | 57.0 | $84 |
| Gemma 3 12BGoogle | 47.4 | $78 |
| GPT-5 nanoOpenAI | 43.5 | $54 |
| gpt-oss-20bOpenAI | 34.5 | $32 |