Skip to main content

Cheapest Frontier-Grade LLM APIs in 2026

The most capable LLMs available — GPT-4o, Claude, Gemini Pro, and other frontier models. Compare pricing across providers.

What is the cheapest frontier-grade LLM?

The cheapest frontier-grade LLM API is DeepSeek R1 0528 Qwen3 8B via Novita AI at $0.068 per million tokens. 148 frontier models are compared across all providers.

Frontier-Grade Models by Price

148 models
#ModelInputOutput
12DeepSeek: DeepSeek V3.2 Exp
Open
$0.210$0.320
16DeepSeek V3.2
$0.260$0.380
33DeepSeek V3.2 Speciale
$0.480$1.44
38accounts/fireworks/models/deepseek-v3p2
$0.560$1.68
32GPT 5.1 Codex Mini
$0.250$2.00
31GPT 5 Mini
$0.250$2.00
28Nex AGI: DeepSeek V3.1 Nex N1
$0.324$1.20
1DeepSeek R1 0528 Qwen3 8B
$0.060$0.090
9DeepSeek R1 0528 Distill Qwen3 8B
$0.200$0.200
22DeepSeek: DeepSeek V3.1 Terminus
Open
$0.210$0.790
23DeepSeek V3.1
$0.210$0.790
24DeepSeek: DeepSeek V3.1 Terminus (exacto)
Open
$0.252$0.948
39accounts/fireworks/models/deepseek-v3p1
$0.560$1.68
48DeepSeek: R1 0528
Open
$0.500$2.15
36mistral-large-3
$0.500$1.50
14Gemini 2.5 Flash Preview 04 17
$0.150$0.600
40Gemini 2.5 Flash Preview 05 20
$0.300$2.50
41Google: Gemini 2.5 Flash Image (Nano Banana)
$0.300$2.50
42Gemini 2.5 Flash Native Audio Preview 12 2025
$0.300$2.50
43Gemini 2.5 Flash Native Audio Latest
$0.300$2.50
44Google: Gemini 2.5 Flash
$0.300$2.50
45Gemini 2.5 Flash Native Audio Preview 09 2025
$0.300$2.50
46Gemini 2.5 Flash Preview 09 2025
$0.300$2.50
47Gemini 2.5 Flash Preview TTS
$0.300$2.50
4Llama 4 Scout 17B 16e Instruct
$0.080$0.300
13Llama 4 Maverick 17B 128e Instruct Fp8
$0.150$0.600
17Llama 4 Maverick 17B 128E Instruct
$0.200$0.600
19Grok 3 Mini Latest
$0.300$0.500
20Grok 3 Mini
$0.300$0.500
21Grok 3 Mini Beta
$0.300$0.500
2DeepSeek R1 Distill Qwen 1.5b
$0.100$0.100
5DeepSeek R1 Distill Qwen 14B
$0.150$0.150
10DeepSeek R1 Distill Llama 8B
$0.200$0.200
11DeepSeek R1 Distill Qwen 7B
$0.200$0.200
15DeepSeek: R1 Distill Qwen 32B
32BOpen
$0.270$0.270
26TNG: DeepSeek R1T2 Chimera
$0.300$1.02
34DeepSeek: R1 Distill Llama 70B
70BOpen
$0.700$0.800
49DeepSeek R1 Basic
$0.550$2.19
50DeepSeek R1
$0.550$2.19
18DeepSeek V3 0324
$0.200$0.770
25DeepSeek V3
$0.320$0.890
6Gemini 2.0 Flash Preview Image Generation
$0.100$0.400
7Gemini 2.0 Flash
$0.100$0.400
8Google: Gemini 2.0 Flash
$0.100$0.400
30Gemini 2.0 Flash Live 001
$0.350$1.50
3accounts/fireworks/models/llama-v3p1-405b-instruct-long
$0.100$0.100
37Meta Llama 3.1 405B Instruct
405B
$0.800$0.800
35Mistral Large 3 675B Instruct
$0.500$1.50
27meta-llama/Llama-4-Maverick-17B-128E-Instruct-Turbo
$0.500$0.500
29deepseek/deepseek-v3-turbo
$0.400$1.30

* Prices include ~20% OpenRouter platform fee (verified against actual billing).

Frequently Asked Questions

The cheapest frontier-grade LLM API is DeepSeek R1 0528 Qwen3 8B via Novita AI at $0.068 per million tokens (blended rate). There are 148 frontier models available.