Data-driven ranking

Cheapest AI models for a balanced token workload

Compare the cost of 1M uncached input plus 1M output tokens across verified USD API quotes.

30 eligible offerings · Latest source verification: 2026-09-05T17:44:19.971Z

How this ranking works

Only verified USD input/output quotes qualify. Sorted by the total for 1M uncached input plus 1M output tokens, ascending. This is an aggregate workload split across requests, not a 2M-token request. Equal values share a rank; slug order only breaks display ties. Affiliate status never affects order.

Coverage is our database, not every API on the market. Default non-batch tiers apply, including DeepSeek peak rates and fixed OpenRouter backend quotes. Cache, long-context surcharges, optional tools, hosting conditions and tokenizers can change your bill. Compare all tiers and source terms before choosing.

Scroll to compare all columns →

Cheapest AI models for a balanced token workload — verified USD quotes
Rank	Model / Provider	Input / 1M	Output / 1M	1M input + 1M output	Last verified / Source
1	OpenAI: gpt-oss-20b · OpenRouter / deepinfra/bf16OpenRouter	$0.03	$0.14	$0.17	2026-09-05Official source
2	OpenAI: gpt-oss-120b · OpenRouter / deepinfra/bf16OpenRouter	$0.037	$0.17	$0.207	2026-09-05Official source
3	Meta: Llama 3.3 70B Instruct · OpenRouter / deepinfra/turboOpenRouter	$0.10	$0.32	$0.42	2026-09-05Official source
4	OpenAI GPT OSS 120BFireworks AI	$0.15	$0.60	$0.75	2026-09-05Official source
4	Mistral Small 4Mistral AI	$0.15	$0.60	$0.75	2026-09-05Official source
6	CodestralMistral AI	$0.30	$0.90	$1.20	2026-09-05Official source
7	GPT-5.6 LunaOpenAI	$0.20	$1.20	$1.40	2026-09-04Official source
8	MiniMax M3Fireworks AI	$0.30	$1.20	$1.50	2026-09-05Official source
8	MiniMax M3 · Together AITogether AI	$0.30	$1.20	$1.50	2026-09-05Official source
10	Gemini 3.1 Flash-LiteGoogle	$0.25	$1.50	$1.75	2026-09-04Official source
11	DeepSeek V4 FlashDeepSeek	$0.44	$1.32	$1.76	2026-09-04Official source
12	Muse Glimmer 30BFireworks AI	$0.35	$1.50	$1.85	2026-09-05Official source
13	Aya Expanse 32BCohere	$0.50	$1.50	$2.00	2026-09-05Official source
13	Qwen 3.7 PlusFireworks AI	$0.40	$1.60	$2.00	2026-09-05Official source
13	Mistral Large 3Mistral AI	$0.50	$1.50	$2.00	2026-09-05Official source
16	Grok 4.3xAI	$1.25	$2.50	$3.75	2026-09-04Official source
17	Kimi K2.7 CodeFireworks AI	$0.95	$4.00	$4.95	2026-09-05Official source
18	DeepSeek V4 ProDeepSeek	$1.32	$3.96	$5.28	2026-09-04Official source
19	GLM 5.3Fireworks AI	$1.40	$4.40	$5.80	2026-09-05Official source
20	Qwen 3.8 MaxFireworks AI	$2.00	$6.00	$8.00	2026-09-05Official source
20	Grok 4.5xAI	$2.00	$6.00	$8.00	2026-09-04Official source
20	Grok 4.6xAI	$2.00	$6.00	$8.00	2026-09-04Official source
20	Qwen3.8-2.4T-A95B · Together AITogether AI	$2.00	$6.00	$8.00	2026-09-05Official source
24	Mistral Medium 3.5Mistral AI	$1.50	$7.50	$9.00	2026-09-05Official source
25	Claude Sonnet 5Anthropic	$2.00	$10.00	$12.00	2026-09-04Official source
26	GPT-5.6 TerraOpenAI	$2.00	$12.00	$14.00	2026-09-04Official source
27	Kimi K3Fireworks AI	$3.00	$15.00	$18.00	2026-09-05Official source
27	Kimi K3 · Together AITogether AI	$3.00	$15.00	$18.00	2026-09-05Official source
29	GPT-5.6 SolOpenAI	$4.00	$20.00	$24.00	2026-09-04Official source
30	GPT-6 AstraOpenAI	$10.00	$50.00	$60.00	2026-09-04Official source

Your workload may change the ranking

Set your actual input and output volumes to compare total costs.

Open cost calculator Compare models

Cheapest AI models for a balanced token workload

How this ranking works

Your workload may change the ranking

Related comparisons