Best value Gemma 4 31B API providers

33 providers host Gemma 4 31B. Best value (SWE-bench score divided by blended cost) is OpenInference at $0.13/MTok blended.

What would Gemma 4 31B cost you?

OpenInference is 88% cheaper than Cerebras at this workload.
Input tokens / month150M
Output tokens / month30M

Projected monthly cost = (input price × 150M) + (output price × 30M). Drag the sliders to match your actual workload; the chart re-ranks live.

#HostContextInput $/MTokOutput $/MTokBlendedUptime 30mQuant
1OpenInference262k$0.08$0.35$0.1398.79%bf16
2DeepInfra262k$0.09$0.34$0.1398.90%fp4
3DeepInfra262k$0.09$0.34$0.1399.85%fp4
4CoreWeave262k$0.10$0.34$0.1498.20%fp4
5CoreWeave262k$0.10$0.34$0.1499.21%fp4
6WandB262k$0.12$0.35$0.1699.91%bf16
7Venice256k$0.12$0.36$0.1699.70%bf16
8Venice256k$0.12$0.36$0.1699.37%bf16
9Chutes131k$0.12$0.37$0.1699.04%fp4
10Chutes131k$0.12$0.37$0.1698.54%fp4
11SiliconFlow262k$0.13$0.40$0.1789.45%fp8
12SiliconFlow262k$0.13$0.40$0.1790.70%fp8
13AkashML131k$0.14$0.40$0.1891.10%fp8
14Morph175k$0.14$0.40$0.1892.30%fp4
15Crusoe262k$0.14$0.40$0.1899.19%
16Friendli262k$0.14$0.40$0.1899.97%
17Novita262k$0.14$0.40$0.1898.61%bf16
18Crusoe262k$0.14$0.40$0.1899.80%
19Friendli262k$0.14$0.40$0.1899.94%
20Novita262k$0.14$0.40$0.1899.92%bf16
21Parasail262k$0.15$0.40$0.1999.46%fp8
22Parasail262k$0.15$0.40$0.1999.92%fp8
23Phala262k$0.15$0.46$0.2099.92%
24Phala262k$0.15$0.46$0.2098.91%
25Ambient66k$0.20$0.80$0.3099.48%
26Together262k$0.28$0.86$0.3895.05%
27Together262k$0.28$0.86$0.38100.00%
28SambaNova131k$0.38$1.15$0.51100.00%
29SambaNova131k$0.38$1.15$0.5199.74%
30ModelRun262k$0.75$1.00$0.7999.98%fp4
31ModelRun262k$0.75$1.00$0.7999.97%fp4
32Cerebras131k$0.99$1.49$1.07100.00%fp16
33Cerebras131k$0.99$1.49$1.07100.00%fp16

Other rankings for Gemma 4 31B