Ideal para: Cheap Bulk Workloads
Best LLM for Cheap Bulk Workloads
Ranked primarily on input and output $/1M with a benchmark floor so you do not ship junk at volume.
Atualizado May 2026. Top 3 deste mês: MiMo-V2-Flash, Hunyuan A13B Instruct, Phi 4.
Podium
This month’s top three.
- 1MiMo-V2-FlashXiaomi
- Input / 1M
- $0.09
- Output / 1M
- $0.29
- Context
- 262,144
- 2Hunyuan A13B InstructTencent
- Input / 1M
- $0.14
- Output / 1M
- $0.57
- Context
- 131,072
- 3Phi 4Microsoft
- Input / 1M
- $0.07
- Output / 1M
- $0.14
- Context
- 16,384
Como classificamos
Weights tuned for cheap bulk workloads.
Some workloads are massive but forgiving — classification, tagging, summarization, PII scrubbing. The question is: what is the cheapest model that still clears the quality floor? We weight price dominantly here but set a benchmark floor so the recommendation is not useless.
Our full methodology is published on the página de metodologia.
Pilares e pesos:
- input price50%
- output price30%
- MMLU20%
Full ranking
Modelos no topo
| Posição | Modelo | Fornecedor | Entrada $/1M | Saída $/1M | Contexto |
|---|---|---|---|---|---|
| 1 | MiMo-V2-Flash | Xiaomi | $0.09 | $0.29 | 262,144 |
| 2 | Hunyuan A13B Instruct | Tencent | $0.14 | $0.57 | 131,072 |
| 3 | Phi 4 | Microsoft | $0.07 | $0.14 | 16,384 |
| 4 | Llama 3.3 70B Instruct | Meta | $0.12 | $0.38 | 131,072 |
| 5 | Qwen2.5 72B Instruct | Qwen | $0.12 | $0.39 | 32,768 |
| 6 | Gemma 4 31B | $0.13 | $0.38 | 262,144 | |
| 7 | Olmo 3 32B Think | Allen AI | $0.15 | $0.50 | 65,536 |
| 8 | Qwen3 32B | Qwen | $0.08 | $0.24 | 40,960 |
| 9 | Llama 3.1 70B Instruct | Meta | $0.40 | $0.40 | 131,072 |
| 10 | Qwen3.5-9B | Qwen | $0.10 | $0.15 | 262,144 |
Field notes
Dicas para cheap bulk workloads
- 01
Use batch pricing aggressively. 50%+ discounts are common.
- 02
Use cached-input pricing for repeating preambles.
- 03
A cheaper model with a short retry loop often beats a more expensive model one-shot.
FAQ
Perguntas frequentes
The questions teams ask before picking a model for cheap bulk workloads.
Get instant answers from our AI agent
As of May 2026, our weighted top 3 cheapest-but-capable are MiMo-V2-Flash, Hunyuan A13B Instruct, Phi 4.
Often, yes. Providers offer 30–50% discounts on async batch endpoints in exchange for up-to-24h latency.
When its lower accuracy causes retries, downstream fixup, or human review. Always measure end-to-end dollars-per-correct-output, not dollars-per-token.