Cheapest LLM APIs You Can Call Right Now (2026)

As of 2026-11-04, LLM Reference's top pick for low-cost API calls is Granite 4.0 H Micro by IBM Research (Input $/1M $0.017, $0.112/1M output tokens), chosen from models that meet the podium eligibility rules. Second pick: Llama 3 8B Instruct by AI at Meta.

Last refreshed 2026-11-04. Next refresh: weekly.

The cheapest LLM APIs you can call today, ranked by input price with a quality score beside each so you see the trade-off.

Ranking rule. Use this page when token price is the first constraint and quality still matters. The models below exclude zero-dollar tiers and surface a quality watermark beside listed input prices.

Related intent. Need no-cost options? Compare the free-model leaderboard separately from paid API pricing.

Verdict

Use Granite 4.0 H Micro for low-cost API calls today.

Llama 3 8B Instruct is the runner-up: $0.017 vs $0.050 on Input $/1M.

Researched 45d agoWhy this pick Methodology

1stTop pick

Researched 45d ago

Granite 4.0 H Micro

Input $/1M: $0.017
Output (from): $0.112 / 1M

Try on provider Model detail Compare

2ndShortlist

Researched 28d ago

Llama 3 8B Instruct

Input $/1M: $0.050
Output (from): $0.200 / 1M

Try on provider Model detail Compare

3rdShortlist

Researched 28d ago

Qwen1.5-7B

Input $/1M: $0.050
Output (from): $0.250 / 1M

Try on provider Model detail Compare

How we rank

The cheapest LLM APIs list is a price board. Each model shows a quality score, but the score does not change the order.

Eligibility — Chat-completion APIs that are not deprecated, with an input price and an output price above $0 (free models are on the free page). We exclude embedding, ranking, classification, moderation, extraction, safety, medical, image, vision, video, audio and OCR specialists, as well as decision, routing and classifier models. Each model needs an MMLU or GPQA Diamond score.
Primary ranking — Ascending by the lowest listed input price per 1M tokens across eligible provider routes.
Quality watermark — Each model shows its MMLU score, or its GPQA Diamond score if MMLU is unavailable, as a capability check, but the score does not change cheap-page order.
Variant collapse — We keep one row per model family (`familySlug` + parameter tier). When headline scores tie within ±0.5 pt (±10 Elo on Chatbot Arena), we pick the canonical SKU by lowest tracked input price, then GA over preview or limited access, then newest `release`. A folded sibling within the benchmark noise band can show a "Tied within margin" chip on that score cell.
Podium eligibility — Preview, restricted, retired and retiring models never fill a podium slot. Preview means preview or invite-only access, or a model that is announced or pre-release. Restricted means restricted, partner, private or otherwise unavailable access. Retired means the model is past its retirement date or every provider route is retired. Retiring means the model, or any one of its provider routes, has a retirement date within the next 60 days. These models can still appear in the table below the podium.
Coverage caveat — Models without an MMLU or GPQA Diamond score are left off this list. Pair this list with the coding leaderboard or another task board for quality guardrails.

#	Model	Quality watermark	Context	Input $/1M	Output $/1M
1	Granite 4.0 H Micro Tools Quality watermark: MMLU 67.43%	MMLU 67.43%	131K	$0.02	$0.11
2	Mistral NeMo Instruct (2407) Quality watermark: MMLU 81.5%	MMLU 81.5%	128K	$0.02	$0.03
3	Llama 3.2 1B Instruct Verify pricing Quality watermark: MMLU 49.3%	MMLU 49.3%	128K	$0.03	$0.10
4	Qwen3-8B Quality watermark: GPQA Diamond 58.9%	GPQA Diamond 58.9%	128K	$0.04	$0.14
5	Gemma 7B Instruct Verify pricing Quality watermark: MMLU 75.3%	MMLU 75.3%	8K	$0.05	$0.20
6	Llama 3 8B Instruct Quality watermark: MMLU 76.9%	MMLU 76.9%	8K	$0.05	$0.20
7	Phi-2 Quality watermark: MMLU 68.3%	MMLU 68.3%	2K	$0.05	$0.07
8	Phi-3 Mini 128K Verify pricing Quality watermark: MMLU 76.5%	MMLU 76.5%	128K	$0.05	$0.10
9	Qwen1.5-7B Quality watermark: MMLU 69.4%	MMLU 69.4%	32K	$0.05	$0.25
10	Qwen2.5-7B-Instruct Verify pricing Quality watermark: MMLU 81.2%	MMLU 81.2%	128K	$0.05	$0.05
11	Zephyr 7B Beta Quality watermark: MMLU 71.4%	MMLU 71.4%	—	$0.05	$0.20
12	LFM2.5-24B-A2B ReasoningTools Quality watermark: GPQA Diamond 74.42%	GPQA Diamond 74.42%	128K	$0.05	$0.70
13	Qwen3-30B-A3B Quality watermark: GPQA Diamond 61.6%	GPQA Diamond 61.6%	128K	$0.05	$0.30
14	Gemma 2 9B Verify pricing Quality watermark: MMLU 71.5%	MMLU 71.5%	8K	$0.06	$0.18
15	Nemotron-Nano-9B-v2 Quality watermark: GPQA Diamond 64%	GPQA Diamond 64%	128K	$0.06	$0.23
16	Granite 4.2 8B ReasoningTools Quality watermark: GPQA Diamond 64.14%	GPQA Diamond 64.14%	128K	$0.06	$0.25
17	Qwen3.5-Flash Vision Quality watermark: GPQA Diamond 84.2%	GPQA Diamond 84.2%	1M	$0.07	$0.26
18	Hunyuan Hy3 Preview PreviewReasoningToolsVerify pricing Quality watermark: GPQA Diamond 87.2%	GPQA Diamond 87.2%	256K	$0.07	$0.26
19	Phi-4 14B Verify pricing Quality watermark: MMLU 84.8%	MMLU 84.8%	16K	$0.07	$0.14
20	Qwen2-1.5B Quality watermark: MMLU 69.7%	MMLU 69.7%	—	$0.07	$0.07

Honorable mentions

Next seats in this ranking. Lines below are from each model's stored description in LLMReference seed data—spot-check the model page before relying on a capability claim.

#4Qwen2.5-7B-InstructVerify pricing
Instruction-tuned 7B variant combining strong reasoning with real-time inference on single GPUs, ideal for developer tools and vision applications.
$0.050
Input $/1M
#5Zephyr 7B Beta
Zephyr 7B Beta is a 7-billion parameter large language model, fine-tuned from the Mistral-7B-v0.1 model. It is tailored to serve as an effective virtual assistant, performing well in generating human-like responses. The model's training involved Direct Preference Optimization (DPO) on a combination of publicly available and synthetic datasets, achieving strong performance on benchmarks like MT-Bench and AlpacaEval, especially for conversational tasks. However, its complexity falls short when compared to proprietary models, especially in tasks involving coding and mathematics. A notable limitation is its insufficient alignment to human safety preferences and the absence of in-the-loop filtering to prevent problematic outputs. Zephyr 7B Beta is English-based and carries an MIT license.
$0.050
Input $/1M
#6LFM2.5-24B-A2B
LFM2.5-24B-A2B is Liquid AI's largest LFM2.5 model: a 24B-parameter mixture-of-experts reasoning model with about 2B active parameters per token, built for fast on-device agents. It keeps the LFM2 hybrid of gated short convolutions and GQA (40 layers, 64 experts, top-4 routing), extends context from 32,768 to 131,072 tokens, doubles the vocabulary to 128K, and scales pre-training from 17T to 33T tokens. Unlike LFM2-24B-A2B, it is reasoning-only (explicit chain of thought before the final answer) and is post-trained for tool calling and multi-step agent harnesses. Open weights under LFM Open License v1.0 on Hugging Face; also on Liquid Console, OpenRouter, and Vercel AI Gateway.
$0.050
Input $/1M

Compare Top Picks

Side-by-side comparison of the top picks by price, benchmark, and API access.

Granite 4.0 H Micro vs Llama 3 8B Instruct Granite 4.0 H Micro vs Qwen1.5-7B Llama 3 8B Instruct vs Llama 3.2 1B Instruct Llama 3 8B Instruct vs Gemma 7B Instruct Llama 3 8B Instruct vs Qwen2.5-7B-Instruct Llama 3.2 1B Instruct vs Gemma 7B Instruct

Browse Other Categories

Best LLMs for Code Generation Best LLMs for RAG Best LLMs for Agents vs Agent Products (2026)Best LLMs for Classification Best Open Source LLMs Best Multimodal / Vision LLMs Best LLM for Translation in 2026 Best AI Image Models in 2026 Best AI Video Models in 2026 Best LLMs for Reasoning & Math Best Small Language Models (SLMs)Best LLMs for JSON / Tool use Best Long Context LLMs Best Mainstream LLM APIs, Ranked Best LLMs for Enterprise Best Free LLMs You Can Use Right Now Best LLMs for Writing Best LLMs for Marketing Best LLMs for Customer Support

Frequently asked questions

What is the best LLM for low-cost API calls?

LLM Reference's top pick for low-cost API calls as of 2026-11-04 is Granite 4.0 H Micro by IBM Research, with Input $/1M $0.017. Picks come from models that meet the podium eligibility rules, so the top pick may not lead the full table. Second pick: Llama 3 8B Instruct by AI at Meta (Input $/1M $0.050).

Which ranked LLM for low-cost API calls has the lowest output price?

Among the ranked models shown on this page, Mistral NeMo Instruct (2407) by MistralAI has the lowest listed output price: $0.030/1M output tokens, at rank 2.

What is the best open-weight LLM for low-cost API calls?

Granite 4.0 H Micro by IBM Research is the highest-ranked open-weight model shown on this page (rank 1, Input $/1M $0.017). License: Apache 2.0.

How is the low-cost API calls ranking calculated?

The cheapest LLM APIs list is a price board. Each model shows a quality score, but the score does not change the order. Preview, restricted, retired and retiring models never fill a podium slot. Preview means preview or invite-only access, or a model that is announced or pre-release. Restricted means restricted, partner, private or otherwise unavailable access. Retired means the model is past its retirement date or every provider route is retired. Retiring means the model, or any one of its provider routes, has a retirement date within the next 60 days. These models can still appear in the table below the podium.

How often is this page updated?

Weekly. Last refreshed 2026-11-04.