$0.002 in / $0.002 out
- Context: 33K
- Providers: 5
- Tool calling
- Structured output
- Open weights
Best AI models · 2026-09-28
Cheapest LLM APIs in 2026
Lowest-priced text models from established vendors.
How we picked these
- Sorted by sum of input + output price per 1M tokens.
- Models with $0 placeholder pricing (Github Copilot, free-tier rebroadcasts) are excluded — Unknown ≠ free.
- Only text models with valid context windows are included.
Top 10 picks
$0.009 in / Unknown out
- Context: 256
- Providers: 1
- Open weights
$0.009 in / Unknown out
- Context: 512
- Providers: 1
- Open weights
$0.010 in / Unknown out
- Context: 33K
- Providers: 7
$0.010 in / Unknown out
- Context: 33K
- Providers: 4
$0.010 in / Unknown out
- Context: 8K
- Providers: 1
- Open weights
$0.010 in / $0.010 out
- Context: 60K
- Providers: 6
- Open weights
$0.010 in / $0.010 out
- Context: 33K
- Providers: 3
$0.010 in / $0.010 out
- Context: 512
- Providers: 2
- Open weights
$0.010 in / $0.010 out
- Context: 512
- Providers: 2
- Open weights
Recommended stack by tier
Same shortlist sliced four ways — pick the tier that matches your budget and constraints.
Budget
Mistral$0.002 in / $0.002 out · 33K ctx
Lowest total per-1M-token cost in this list ($0.00).
Lowest-cost option that still meets the use case. Pick this when you have high volume or strict unit-economics.
Balanced
digitalocean$0.010 in / Unknown out · 8K ctx
Median price ($0.01) — typically the safest default.
Good-enough quality at a mid-tier price. The default choice for most production apps.
Premium
Meta$0.010 in / $0.010 out · 512 ctx
Highest-priced pick in the list ($0.02) — usually the flagship.
Highest-capability model in this list. Pick when accuracy or reasoning matters more than cost.
Open-weight
No fit in this list
Open weights — self-host on your own GPUs, fine-tune on private data, run offline. Pricing here reflects the cheapest API host.
Frequently asked questions
Which AI model is the best for the lowest possible cost in 2026?
Right now we put Voxtral Small 24B 2507 from Mistral at the top, primarily because its sum of input + output per-million-token cost is the lowest among models that are not placeholder-priced. Rankings are recomputed from live model metadata — see "How we picked these" above for the exact rule.
What is the cheapest option in this list?
Voxtral Small 24B 2507 (Mistral) is the lowest-priced pick at $0.002 per 1M input tokens and $0.002 per 1M output tokens. Costs from other entries scale up from there.
How are these rankings generated?
Each pick comes from a programmatic rule defined in our use-case-rules config: a hard filter (e.g. tool calling required, context ≥ 100K) plus a numeric score combining capability, context window and price. We never hand-curate the order, but we do hand-curate the rule. Underlying model metadata is refreshed daily from a normalised canonical catalogue.
How often is this page updated?
The underlying model data is refreshed once per day, and the static page is rebuilt when the data changes. The 'Last updated' date below shows the most recent rebuild.
Are 'Unknown' priced models excluded?
Yes. We only rank models with a published per-million-token price. 'Unknown' here means the provider does not publish a public rate card — it is not the same as 'free', so showing them at $0 would mislead.
Explore more
Top picks · model details
- Voxtral Small 24B 2507$0.00 in / $0.00 out
- All-MiniLM-L6-v2$0.01 in / $0.00 out
- Multi-QA-mpnet-base-dot-v1$0.01 in / $0.00 out
- Qwen3 Embedding 8B$0.01 in / $0.00 out
- Qwen3 Embedding 4B$0.01 in / $0.00 out
Other best-of lists
Browse by capability
Vendors in this list
Last updated:
Prices in USD per 1M tokens. Unknown means the provider does not publish per-token pricing.
Pricing and capabilities are refreshed daily and reconciled against each provider's official documentation. Always verify critical production decisions with the provider directly.