Where to run open-weight models for free or cheap in 2026

tldr: October 2026 brought four open-weight releases that matter for coding: Muse Spark Open, DeepSeek V4.1 Flash, MiniMax M3.1 Flash, and Kimi K3.1 Pro, whose weights are due by 4 November. Free access to models like these exists but is small, and Google and some OpenRouter providers train on free-tier prompts. Most paid hosts charge the lab's list price, so off-peak hours, cache-read rates, and coding plans move your bill more than the host does. If you want the models to ship code rather than answer API calls, Buildful runs them as a background agent for $10 a month.

Which open-weight models are worth running right now?

Three of these shipped between 15 and 19 October 2026, and a fourth has weights due by 4 November. Prices are per million tokens, input and output.

Pick DeepSeek V4.1 Flash if you want a model with no conditions attached. The 1017 snapshot is MIT-licensed, has a one-million-token window, and costs $0.30 and $1.20 at every host that serves it, or half that on DeepSeek's own API off-peak.

The other October releases put conditions in the license. Meta's Muse Spark Open has 816 billion parameters, 42 billion of them active, and takes text, image, video, and audio input. Its license makes every product built on it display "Built with Muse Spark Open", and a model-as-a-service business past $20 million a year needs a separate deal with Meta. MiniMax pitches M3.1 Flash at coding, agents, and computer use for $0.15 and $0.60. Its license asks for a "Built with MiniMax M3.1-Flash" notice and a one-time email to MiniMax before commercial use, and written permission once your revenue passes $20 million.

Kimi K3.1 Pro is the largest of the four and the least settled. Moonshot calls it its "most capable 3T-class open-weight model" and bills it flat, with no surcharge for filling its one-million-token window. The weights aren't public yet. The quickstart promises them by 4 November, and until then Together, Fireworks, DeepInfra, Ollama, and Cloudflare don't serve it. If its license follows Kimi K3's, expect revenue gates there too.

Of the models that shipped before October, Z.ai's own card calls GLM-5.3 "the most capable open-weights model for coding". Its license asks for a security review only from API businesses past $10 billion in revenue, which leaves out almost everyone reading this. Qwen3.8-27B is the small, permissive pick, a compact dense model under Apache 2.0.

Which free tiers are actually free?

Every figure below comes from the provider's own pricing or limits page, read on 27 October 2026.

GitHub Models is missing from that table on purpose. GitHub fully retired it on 30 July 2026, including the playground, catalog, and inference API.

Two of these stand out. Gemma 4 on the Gemini API is the only option where a capable open-weight model costs nothing at any volume the rate limiter allows, provided you are comfortable with Google reading your prompts. And OpenRouter's $10 threshold is the best $10 in this article: it lifts the free-model cap from 50 requests a day to 1,000, and you keep the credits for paid models.

OpenRouter's limits on free models. Screenshot: OpenRouter docs (openrouter.ai), 27 October 2026.

OpenRouter's free roster isn't the models from the table above, though. The free collection is led by NVIDIA's Nemotron family, Poolside's Laguna, and Thinking Machines' Inkling, the open-weight model from our fine-tuning guide. DeepSeek, GLM, and Kimi aren't on it. Read the model card before you send anything private: Poolside says it "may use your inputs and outputs to train" when you use Laguna for free, and Liquid's card says prompts to LFM2.5 may be retained for training.

The Cloudflare allowance is easier to judge in tokens. At gpt-oss-120b's paid rates, $0.11 a day buys about 314,000 input tokens or 147,000 output tokens. That covers a hobby project. The sample agent month in the cost section reads about 3.3 million input tokens a day, ten times what the allowance covers.

Does the host change the price?

Mostly, no. Line up the same model across hosts and the rates are identical to the cent, because hosts match the lab's own list rate. Prices are per million tokens, input and output.

Sources: Z.ai, DeepSeek, Together, Fireworks, DeepInfra, Ollama, and Cloudflare. The DeepSeek figures are its peak rates. The Kimi K3 lab price is from our launch coverage, since Moonshot's own pricing table rendered empty on 27 October.

GLM-5.3 and Muse Spark Open cost the same at every host that lists them. Prices read on 27 October 2026.

GLM-5.3 costs the same at five hosts. Muse Spark Open costs the same at all three hosts that list it. So asking which host is cheapest gets you nowhere. Ask who is fastest, who serves the model you need, and how they bill cache reads.

Speed is the one place the hosts separate. Cerebras lists gpt-oss-120b at about 3,000 tokens a second and charges $0.35 and $0.75 for it. Groq lists the same model at 500 tokens a second for $0.15 and $0.60. You pay a little more than double on input to go six times faster, which is a good trade for an agent loop that waits on every response.

DeepInfra is the exception on price. Its DeepSeek V4 Pro output costs 34% less than everyone else's, and it still serves the July snapshot of V4 Flash at $0.06 and $0.18. On agent traffic, that output discount buys less than it seems.

What actually cuts the bill?

Take one illustrative month of coding-agent traffic: 100 million input tokens, 80 million of them cache hits, and 3 million output tokens. Agents reread the same repository context on every turn, so a heavy input side and a high cache-hit rate is the normal shape.

One month of coding-agent traffic (100M input tokens, 80M of them cached, and 3M output tokens) priced on each setup.

Three things in that table move the bill more than the host does.

Off-peak hours halve DeepSeek. DeepSeek's own API charges half price off-peak. Peak runs 01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays, which is 35 of the week's 168 hours. The other 79% is off-peak, including a full 9am to 5pm working day on either US coast. For an American team, V4 Pro effectively costs $0.66 and $1.98, which no other host in this article matches. Ollama runs its own off-peak discount on DeepSeek models, on a different window: outside 12:00 to 18:00 UTC on weekdays, and all weekend.

DeepSeek's own API charges full price for 35 of the week's 168 hours.

DeepSeek's off-peak and peak rates, with the middle rows trimmed. Screenshot: DeepSeek API Docs (api-docs.deepseek.com), 27 October 2026.

Cache reads decide agent bills. DeepInfra's 34% output discount on V4 Pro lands at exactly the same $41.80 as DeepSeek's peak rate. The reason is cache reads: DeepInfra charges $0.10 per million for them, against DeepSeek's $0.044. Together charges $0.13, which is why the same model costs 16% more there. When 80% of your input is cached, the cache-read column outweighs the output column. Check it before you switch hosts for a headline rate.

The model is the biggest lever of all. V4.1 Flash off-peak costs a quarter of V4 Pro off-peak on this workload, and Kimi K3 costs six times V4 Pro off-peak. If one of the Flash models passes your task, that saving dwarfs everything above. Our LLM API pricing comparison makes the same point for closed models: pick the tier from the task, then compare prices inside it.

Are coding plans cheaper than paying per token?

Yes, by a wide margin, if you stay inside the plan's rules.

Z.ai's GLM Coding Plan starts at $18 a month. Its entry tier, Lite, allows 2,000 credits per five hours and 10,000 a week, and Z.ai publishes the conversion: credits equal input tokens times 6.9, plus cached input times 1.7, plus output times 24, all divided by 10,000. Run the workload above through that formula and it costs 34,600 credits. Lite allows about 43,300 a month. The same tokens at GLM-5.3's API rate cost $62, and at this traffic mix the full Lite allowance is worth about $78 at list price. Off-peak usage (outside 14:00 to 18:00 Singapore time on weekdays) burns credits at half rate. Z.ai's FAQ lists the catches. The plan works only with supported tools such as Claude Code, Codex, OpenCode, and Cline, subscriptions are non-refundable, and the five-hour window caps bursts at about 61 requests of 30,000 uncached input and 5,000 output tokens.

OpenCode Go costs $10 a month and covers 30 models, open and closed, including Kimi K3.1 Pro, MiniMax M3.1 Flash, DeepSeek V4.1 Flash, GLM-5.3, and Qwen3.8 Max. Each model has a separate monthly dollar budget, from $80 of DeepSeek V4.1 Flash and $20 of MiniMax M3.1 Flash down to $15 each of GLM-5.3 and Kimi K3.1 Pro. Any five-hour window can use 20% of a model's budget and any week 50%. It works with other coding agents and exposes OpenAI-compatible endpoints, though OpenCode says it monitors traffic for abuse and wants agent-shaped requests.

Per-model monthly limits on OpenCode Go. Screenshot: OpenCode docs (opencode.ai), 27 October 2026.

Ollama sells the same idea with less fine print. Pro costs $20 a month and includes $60 of usage credits at its normal per-token rates. Max costs $100 and includes $300. That's three times face value on any cloud model Ollama hosts, Muse Spark Open, GLM-5.3, and DeepSeek V4 Pro included. Unused credit does not roll over.

MiniMax and Moonshot sell their own coding plans too, but neither publishes a token figure. MiniMax's M Plan describes its Explore and Build tiers as 3 and 7.5 times the usage of Go, and Kimi Code meters on a rolling five-hour window.

What if you want the code written, not an API?

Every option above still leaves you to wire up the agent, the sandbox, the test run, and the review. That's fine if you already live in Claude Code. If you have a GitHub repo and a backlog and want a Claude Code alternative with nothing to assemble, look at Buildful.

Buildful is an AI coding agent: you describe a change, and it builds it, tests it in a real browser, and opens a pull request on your GitHub repo. It's a background coding agent that runs open-weight models in the cloud, so there's no GPU, no API key, and nothing running on your laptop. It works like a small AI software factory: a task goes in, a tested pull request comes out, and you can close the tab while it works.

You start a task in the browser, Slack, or Linear. Buildful works on your repo in a private cloud sandbox that is deleted when the task ends, runs your lint, type checks, and tests, fixes what fails, and tests UI changes in a real browser with a recording. What comes back is a pull request with the recording attached. Nothing reaches your main branch until you merge it.

By default it picks an open-weight model for each task. On a paid plan you can choose one yourself, and the picker already carries the October releases: Muse Spark Open, DeepSeek V4.1 Flash, MiniMax M3.1 Flash, and Kimi K3.1 Pro, alongside GLM-5.3, GLM-5.3-Flash, and Kimi K3. Frontier models are there too, on prepaid credit, for the tasks that need them.

The open-weight models you can pick on Buildful's paid plans. Screenshot: Buildful (buildful.ai), 27 October 2026.

Pricing is flat. Pro costs $10 a month for 1,000 tasks, which works out to about a cent a task, and Team costs $40 for 3,000 tasks shared by everyone on it. If you lean on the models you pick yourself, the Buildful Pass adds $80 of usage a month on four open-weight models for $30. You can try three tasks free, with no card. Its privacy policy says it does not use your code or task content to train models.

It's the wrong tool if you need an API key for your own product or a base URL to point Claude Code at, because Buildful offers neither.

How do you point Claude Code at an open-weight model?

Z.ai, MiniMax, and Moonshot all run Anthropic-compatible endpoints, so Claude Code needs two environment variables and no other change:

export ANTHROPIC_BASE_URL=https://api.z.ai/api/anthropic
export ANTHROPIC_AUTH_TOKEN=your-zai-api-key
claude

MiniMax's base URL is https://api.minimax.io/anthropic, and Kimi's overseas endpoint is https://api.kimi.ai/coding/. Z.ai's setup page and MiniMax's also show how to map Claude Code's model slots to their models. Once the variables are set, requests stop drawing on your Anthropic plan, so this doubles as a way around Claude Code's usage limits.

For everything else, an OpenAI-compatible client and OpenRouter cover the most ground with one key. OpenRouter doesn't mark up provider rates. It charges 5.5% when you buy credits, so routing $100 of usage through it costs $105.50.

Can you run them locally for free?

If you own the hardware, yes. Ollama's local runtime is MIT-licensed and unmetered. OpenAI says gpt-oss-20b runs in 16 GB of memory, which puts it on a recent laptop, and Qwen3.8-27B is a 27-billion-parameter dense model its card calls deployment-friendly. The October releases are a different matter. Muse Spark Open is 816 billion parameters and Kimi K3's weights alone are 1.4 TB, so self-hosting either means renting or owning a GPU cluster.

Which one should you pick?

FAQ

Is there a completely free API for open-weight models?

Yes, with limits. Gemma 4 is free of charge on the Gemini API, OpenRouter offers more than 25 free models, Groq has a free plan for gpt-oss and Qwen3.8 27B, and Cloudflare gives 10,000 neurons a day. Google uses free-tier prompts to improve its products, and some OpenRouter providers train on free traffic.

Where can I use Kimi K3.1 Pro for coding?

Through Moonshot's own API and Kimi Code plan, OpenRouter, OpenCode Go, and Buildful's model picker on paid plans. As of 27 October 2026, Together, Fireworks, DeepInfra, Ollama, and Cloudflare don't serve it. Moonshot says the full weights will be released by 4 November, after which more hosts should follow.

Is there a Claude Code alternative that runs on open-weight models?

Yes. Buildful is an AI coding agent that runs open-weight models in the cloud: you describe a change, and it builds it, runs your checks, tests UI changes in a real browser, and opens a pull request on your GitHub repo. Pro costs $10 a month for 1,000 tasks, and paid plans let you pick models like Kimi K3.1 Pro, GLM-5.3, and DeepSeek V4.1 Flash yourself. You can try three tasks free, with no card.

Is there a subscription for open-weight coding models?

Several. OpenCode Go costs $10 a month with a per-model budget, Z.ai's GLM Coding Plan starts at $18, and Ollama Pro costs $20 with $60 of credits. Buildful costs $10 a month for 1,000 tasks, and its $30 Pass adds $80 of usage on four open-weight models.

Is OpenRouter more expensive than going direct?

Only by its credit fee. Inference bills at the provider's list price, and OpenRouter adds 5.5% when you buy credits on the standard plan.

Is GitHub Models still free?

No. GitHub retired GitHub Models entirely on 30 July 2026. Its docs point to Azure AI Foundry and GitHub Copilot instead.