# Vals Inference Provider Index v1 (Vals Inference Provider Index)

> Vals AI comparison of the same open-weight model served by different inference providers, against the model creator's own API, on agentic finance, coding and legal tasks.

Canonical page: https://benchlm.ai/benchmarks/vals-inference-provider-index

- Category: [Agentic](/agentic)
- Last updated: October 20, 2026

Evaluation results published by [Vals AI](https://www.vals.ai/benchmarks/inference_provider). Published leaderboard aggregates and evaluation settings only. Dataset contents are not included.

## About Vals Inference Provider Index

- Year: 2026
- Tasks: 165 agentic tasks from three Vals Index benchmarks
- Format: Accuracy, cost per task and speed by model and provider
- Difficulty: Long tool-using agent work in finance, coding and law
- Paper: [Vals Inference Provider Index leaderboard and methodology](https://www.vals.ai/benchmarks/inference_provider)

Vals published v1 on October 20, 2026 and reruns it every two weeks. DeepSeek V4.1 Flash, Kimi K3 and MiniMax M3 each run on their creator's API and on Together AI, Fireworks and Baseten (MiniMax M3 was not available on Baseten) over a fixed 165-task set: 66 Finance Agent v2, 33 Vibe Code Bench and 66 Legal Research tasks, with the same agent and aligned sampling settings. Vals reports accuracy, cost per task and speed. We mirror the published aggregates as a display-only board, outside overall and category rankings.

Vals Inference Provider Index is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (11 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Kimi K3 · Fireworks](https://www.vals.ai/models/fireworks_kimi-k3) | — | Fireworks AI | 73.85% |
| 2 | [Kimi K3 · Together AI](https://www.vals.ai/models/together_moonshotai/Kimi-K3) | — | Together AI | 73.18% |
| 3 | [DeepSeek V4.1 Flash · Fireworks](https://www.vals.ai/models/fireworks_deepseek-v4p1-flash) | — | Fireworks AI | 72.99% |
| 4 | [Kimi K3 · Moonshot API](https://www.vals.ai/models/kimi_kimi-k3) | max reasoning | Moonshot AI | 71.77% |
| 5 | [DeepSeek V4.1 Flash · DeepSeek API](https://www.vals.ai/models/deepseek_deepseek-v4.1-flash) | high reasoning | DeepSeek | 70.33% |
| 6 | [DeepSeek V4.1 Flash · Baseten](https://www.vals.ai/models/baseten_deepseek-ai/DeepSeek-V4.1-Flash) | — | Baseten | 68.86% |
| 7 | [Kimi K3 · Baseten](https://www.vals.ai/models/baseten_moonshotai/Kimi-K3) | — | Baseten | 67.28% |
| 8 | [DeepSeek V4.1 Flash · Together AI](https://www.vals.ai/models/together_deepseek-ai/DeepSeek-V4.1-Flash) | — | Together AI | 66.39% |
| 9 | [MiniMax M3 · MiniMax API](https://www.vals.ai/models/minimax_MiniMax-M3) | — | MiniMax | 42.88% |
| 10 | [MiniMax M3 · Fireworks](https://www.vals.ai/models/fireworks_minimax-m3) | — | Fireworks AI | 39.35% |
| 11 | [MiniMax M3 · Together AI](https://www.vals.ai/models/together_MiniMaxAI/MiniMax-M3) | — | Together AI | 37.63% |

## FAQ

### What does Vals Inference Provider Index measure?

Vals AI comparison of the same open-weight model served by different inference providers, against the model creator's own API, on agentic finance, coding and legal tasks.

### Which model leads the published Vals Inference Provider Index snapshot?

Kimi K3 · Fireworks currently leads the published Vals Inference Provider Index snapshot with a score of 73.85%.

### How many models are evaluated on Vals Inference Provider Index?

The October 20, 2026 contains 11 AI models.

### Does Vals Inference Provider Index affect BenchLM's overall score?

Not directly. Vals Inference Provider Index is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.