# Cheapest LLM API in October 2026, and the cheapest for coding

Canonical: https://theagenticweekly.com/what-is/cheapest-llm-api · Last verified: 2026-10-27 · Published: 2026-10-27 · Author: Agentic Weekly (operated independently from Texas, US) · Evidence: Evidence: Reported. Prices are vendor list prices from our model records; the coding ranking is the LMArena WebDev crowd-vote snapshot of Oct 22, 2026, which is a third-party preference score, not our test read from the sources listed below on 2026-10-27 (America/Chicago). We did not run these products or models for this page.

> As of Oct 26, 2026, the cheapest language-model API in our data is Ministral 3 14B at $0.20 / $0.20 per million input / output tokens (USD list price, 41 priced models). For coding, the cheapest models in the top 10 of the LMArena WebDev board (Oct 22, 2026) are Claude Sonnet 5.5, GPT-6.1 Sol, Kimi K3.1 Pro and GPT-6 Sol at $2 / $10 per million tokens. Claude Sonnet 5.5 ranks #3. Cheapest is not best: we have not measured quality per dollar.

Method: Prices are vendor list prices in USD per 1M tokens, standard processing; the coding cut uses the LMArena WebDev crowd-vote snapshot of Oct 22, 2026, a third-party preference score. Each fact is copied from the vendor page linked in its row into our model records with the date we read it. Reported means read from that page, not run by us; Tested would mean we ran it and logged a dated run, and no model on this page is Tested. 64 language models are in our data: 41 have a sourced USD list price, 2 are priced in another currency and 21 have no verified price. Only models with a sourced price are listed; we do not estimate or pad.

Data: [models.csv](https://theagenticweekly.com/models.csv) · [models.json](https://theagenticweekly.com/models.json) · [answers.json](https://theagenticweekly.com/answers.json)

## The 8 cheapest language-model APIs (USD)

Ranked by input price plus output price per 1M tokens, standard processing, free tiers ignored. Full list: [LLM pricing comparison](https://theagenticweekly.com/what-is/llm-pricing-comparison).

Table: Eight cheapest LLM APIs by input plus output price, USD per 1M tokens

| Model | Vendor | Input / output | Licence | Source | Read |
|---|---|---|---|---|---|
| [Ministral 3 14B](https://theagenticweekly.com/models/ministral-3-14b) | Mistral AI | $0.20 / $0.20 | Apache 2.0 | [vendor page](https://docs.mistral.ai/models/ministral-3-14b-25-12) | 2026-10-26 |
| [Mercury 3](https://theagenticweekly.com/models/mercury-3) | Inception | $0.10 / $0.35 | Proprietary (API) | [vendor page](https://www.inceptionlabs.ai/blog/introducing-mercury-3) | 2026-10-26 |
| [Claude Haiku 5.5](https://theagenticweekly.com/models/claude-haiku-5-5) | Anthropic | $0.10 / $0.50 | Proprietary (API) | [vendor page](https://platform.claude.com/docs/en/about-claude/models/overview) | 2026-10-26 |
| [GPT-6 Luna](https://theagenticweekly.com/models/gpt-6-luna) | OpenAI | $0.10 / $0.50 | Proprietary (API) | [vendor page](https://developers.openai.com/api/docs/models/gpt-6-luna) | 2026-10-26 |
| [MiniMax M3.1 Flash](https://theagenticweekly.com/models/minimax-m3-1-flash) | MiniMax | $0.15 / $0.60 | MiniMax Community License | [vendor page](https://www.minimax.io/blog/minimax-m31f) | 2026-10-26 |
| [Mistral Small 4](https://theagenticweekly.com/models/mistral-small-4) | Mistral AI | $0.15 / $0.60 | Apache 2.0 | [vendor page](https://docs.mistral.ai/models/mistral-small-4-0-26-03) | 2026-10-26 |
| [GPT-5.6 Luna](https://theagenticweekly.com/models/gpt-5-6-luna) | OpenAI | $0.20 / $1.20 | Proprietary (API) | [vendor page](https://developers.openai.com/api/docs/models/gpt-5.6-luna) | 2026-10-26 |
| [Grok Build 0.1](https://theagenticweekly.com/models/grok-build-0-1) | xAI | $1 / $2 | Proprietary (API) | [vendor page](https://docs.x.ai/docs/models) | 2026-10-26 |

## Cheapest for coding: the top 10 of LMArena WebDev

LMArena WebDev ranks models on front-end web-development tasks by crowd vote. Scores are Elo-style, snapshot Oct 22, 2026. Prices come from our model records where the model has one, otherwise from the leaderboard row.

Table: LMArena WebDev top 10 with list price per 1M tokens

| Rank | Model | Score | Input / output (USD) | Price source |
|---|---|---|---|---|
| 1 | [Claude Opus 5.5 (Max)](https://theagenticweekly.com/models/claude-opus-5-5) | 1807 | $4 / $20 | [vendor page](https://platform.claude.com/docs/en/about-claude/models/overview) |
| 2 | [GPT-6 Astra (Max)](https://theagenticweekly.com/models/gpt-6-astra) | 1774 | $10 / $50 | [vendor page](https://developers.openai.com/api/docs/models/gpt-6-astra) |
| 3 | [Claude Sonnet 5.5 (xHigh)](https://theagenticweekly.com/models/claude-sonnet-5-5) | 1769 | $2 / $10 | [vendor page](https://platform.claude.com/docs/en/about-claude/models/overview) |
| 4 | [GPT-6.1 Sol (Max)](https://theagenticweekly.com/models/gpt-6-1-sol) | 1749 | $2 / $10 | [vendor page](https://developers.openai.com/api/docs/models/gpt-6.1-sol) |
| 5 | Claude Fable 5.1 (Max) | 1742 | $10 / $50 | [LMArena row](https://arena.ai/leaderboard/webdev) |
| 6 | [Muse Wonder Full (Max)](https://theagenticweekly.com/models/muse-wonder-full) | 1726 | N/A | [LMArena row](https://arena.ai/leaderboard/webdev) |
| 7 | [Claude Sonnet 5.5 (High)](https://theagenticweekly.com/models/claude-sonnet-5-5) | 1709 | $2 / $10 | [vendor page](https://platform.claude.com/docs/en/about-claude/models/overview) |
| 8 | [Kimi K3.1 Pro (Max)](https://theagenticweekly.com/models/kimi-k3-1-pro) | 1692 | $2 / $10 | [vendor page](https://www.kimi.ai/resources/kimi-k3-1-pro-pricing) |
| 9 | Claude Opus 5 (Max) | 1691 | $5 / $25 | [LMArena row](https://arena.ai/leaderboard/webdev) |
| 10 | [GPT-6 Sol (Max)](https://theagenticweekly.com/models/gpt-6-sol) | 1681 | $2 / $10 | [vendor page](https://developers.openai.com/api/docs/changelog) |

Just outside the top 10, DeepSeek-V4.1-Flash-20261017 (Max) is #14 at $0.15 / $0.60 per the leaderboard row. We list that model in Chinese yuan on [its model page](https://theagenticweekly.com/models/deepseek-v4-1-flash-1017) and do not convert, so the dollar figure is LMArena's, not DeepSeek's.

## Cheapest open-weights options

- [Ministral 3 14B](https://theagenticweekly.com/models/ministral-3-14b): $0.20 / $0.20 per 1M tokens, licence Apache 2.0.
- [MiniMax M3.1 Flash](https://theagenticweekly.com/models/minimax-m3-1-flash): $0.15 / $0.60 per 1M tokens, licence MiniMax Community License.
- [Mistral Small 4](https://theagenticweekly.com/models/mistral-small-4): $0.15 / $0.60 per 1M tokens, licence Apache 2.0.
- [GLM-5.3](https://theagenticweekly.com/models/glm-5-3): $1.40 / $4.4 per 1M tokens, licence GLM-5.3 License (MIT-style plus a MaaS condition).

Open weights can also run on your own hardware, where the cost is compute, not tokens. See [best open-source LLM](https://theagenticweekly.com/what-is/best-open-source-llm).

## FAQ

### What is the cheapest LLM API?

Ministral 3 14B at $0.20 / $0.20 per 1M input / output tokens, the lowest USD list price among 41 models in our data (read Oct 26, 2026). DeepSeek lists lower yuan prices but we do not convert currencies.

### What is the cheapest LLM for coding?

Among the top 10 on LMArena WebDev (Oct 22, 2026), Claude Sonnet 5.5, GPT-6.1 Sol, Kimi K3.1 Pro and GPT-6 Sol cost the least at $2 / $10 per 1M tokens; Claude Sonnet 5.5 is the best-ranked of them at #3. DeepSeek-V4.1-Flash-20261017 (Max) is #14 at $0.15 / $0.60 per the leaderboard. These are preference scores from a third party, not our tests.

### Is a cheaper model worse?

Not necessarily, and we have not measured it. Price per token ignores how many tokens a model spends on a task and how often it needs a retry. Check a model's own page for vendor claims, labelled as claims.

## Sources

- [LMArena WebDev leaderboard](https://arena.ai/leaderboard/webdev) (fetched 2026-10-22)
- [Mistral AI pricing: docs.mistral.ai](https://docs.mistral.ai/models/ministral-3-14b-25-12) (fetched 2026-10-26)
- [Inception pricing: inceptionlabs.ai](https://www.inceptionlabs.ai/blog/introducing-mercury-3) (fetched 2026-10-26)
- [Anthropic pricing: platform.claude.com](https://platform.claude.com/docs/en/about-claude/models/overview) (fetched 2026-10-26)
- [OpenAI pricing: developers.openai.com](https://developers.openai.com/api/docs/models/gpt-6-luna) (fetched 2026-10-26)
- [MiniMax pricing: minimax.io](https://www.minimax.io/blog/minimax-m31f) (fetched 2026-10-26)
- [Mistral AI pricing: docs.mistral.ai](https://docs.mistral.ai/models/mistral-small-4-0-26-03) (fetched 2026-10-26)
- [OpenAI pricing: developers.openai.com](https://developers.openai.com/api/docs/models/gpt-5.6-luna) (fetched 2026-10-26)
- [xAI pricing: docs.x.ai](https://docs.x.ai/docs/models) (fetched 2026-10-26)

## Related

- [LLM pricing comparison](https://theagenticweekly.com/what-is/llm-pricing-comparison)
- [Best AI model for coding](https://theagenticweekly.com/best/llm-for-coding)
- [Best open-source LLM](https://theagenticweekly.com/what-is/best-open-source-llm)
- [All AI models](https://theagenticweekly.com/models)
- [Best AI model right now](https://theagenticweekly.com/what-is/best-ai-model-right-now)
- [AI model questions answered](https://theagenticweekly.com/what-is/ai-model-questions-answered)
- [GPT-6.1 Terra vs GPT-6 Astra](https://theagenticweekly.com/compare/gpt-6-1-terra-vs-gpt-6-astra)
- [OpenAI IPO: what is confirmed](https://theagenticweekly.com/what-is/openai-ipo)
- [All explainers and answers](https://theagenticweekly.com/what-is)

## Changelog

- 2026-10-27: First published. Facts read from the sources listed below on this date.

## How to cite

Agentic Weekly, "Cheapest LLM API in October 2026, and the cheapest for coding", last updated 2026-10-27, https://theagenticweekly.com/what-is/cheapest-llm-api