Lite
Gemma 4 E2B
Fast and efficient agents built for speed and throughput. Best for quick answers and simple workflows.
model: "lite"
AI API
The convenience of cloud. The benefits of local.
Private, Compliant, Benchmark Leading Agents under one API you never have to update.
AI is great until it hits production.
A demo can work. Put the same agent in front of real traffic and the misses compound, the model names change out from under you, and every slow call stalls the next step.
A 5–10% error rate compounds.
A few bad tool calls stack until the workflow fails and the experience collapses.
Every model swap is a migration.
New weights ship constantly. Each one means a new name, a new eval, and new code.
Slow calls stall the loop.
Every agent step waits on the model, so latency multiplies across the run.
Stable
Implement once.
Changes under the hood do not mean you have to update your API. Send lite, balanced, or frontier. We choose the model behind each tier and keep it current. When the weights change, your code does not.
Balanced
Gemma 4 26B
Well rounded agent balancing speed, cost and intelligence. Best for most use cases.
model: "balanced"
Frontier
Ornith 1.5 122B
The most intelligent models at a higher cost. Best for complex use cases demanding the best results.
model: "frontier"
Embeddings
Search and retrieval. The current embedding model stays behind this name.
model: "midium"
System One
Decider 4B
Choice, score, and noul. The current System One model stays behind this name.
model: "midium"
See the API. Chat uses lite, balanced, frontier. Embeddings and System One use midium.
It is an OpenAI-compatible API. Embeddings and System One stay on their usual calls, with the model set to midium.
Accuracy
Agents are 5–10% more accurate and reliable with Midium.
AI is a probability engine, which means it is probably right most of the time. Midium mixes deterministic logic with these models to improve tool calling by up to 13% compared to other leading providers.
90.2%
Tool call accuracy on a 26B model at 4-bit. Ollama’s best is 87.4%. Parallel tool calls take a 13.5-point lead.
Speed
Built for production throughput with tens of thousands of monthly active users.
Midium has some of the fastest speeds of any provider on OpenRouter, and is the #1 inference engine for speed and throughput on Mac.
Qwen3.8 Flash Nexttuned
Full results- Short coding ask
- 133 tok/s
- 2.0× Ollama · +5% oMLX
- First token, 33k prompt
- 23×
- ~12.4s vs Ollama’s ~285s · oMLX ~12.9s
- Editing a file
- 3.7×
- 268.1 vs 71.8 tok/s · 1.5× oMLX
Qwen3.6 35B A3BMoE
Full results- Short coding ask
- 304 tok/s
- +7% oMLX · 1.7× Ollama · 2.4× LM Studio
- Editing a file
- +39%
- 603.4 tok/s vs oMLX’s 434.1 · 3.4× Ollama · 5.0× LM Studio
- Writing after a 33k codebase
- +3%
- 189.1 tok/s vs oMLX’s 183.3 · 1.6× Ollama · 1.8× LM Studio
Qwen3.8 27Bdense
Full results- Short coding ask
- 117 tok/s
- +24% oMLX · +48% Ollama · 3.7× LM Studio
- Editing a file
- +47%
- 247.6 tok/s vs oMLX’s 168.9 · 3.0× Ollama · 8.1× LM Studio
- Writing after a 33k codebase
- +8%
- 71.4 tok/s vs oMLX’s 66.1 · +40% Ollama · 2.6× LM Studio
Gemma 4 26B A4BMoE
Full results- Short coding ask
- 247 tok/s
- 1.6× oMLX · +33% Ollama · 1.9× LM Studio
- First token, 37k prompt
- 3.1×
- ~9.6s vs Ollama’s ~30s · oMLX ~11.7s · LM Studio ~10.3s
- Editing a file
- 2.9×
- 451.2 tok/s vs Ollama’s 156.2 · 3.7× oMLX · 3.8× LM Studio
Gemma 4 12Bdense
Full results- Short coding ask
- 150 tok/s
- +25% oMLX · +17% Ollama · 2.3× LM Studio
- First token, 37k prompt
- 3.2×
- ~24s vs Ollama’s ~77s · oMLX ~25s · LM Studio ~24s
- Editing a file
- 2.7×
- 280.6 tok/s vs Ollama’s 104.5 · 3.4× oMLX · 4.7× LM Studio
Gemma 4 31Bdense
Full results- Short coding ask
- 86 tok/s
- +33% oMLX · +20% Ollama · 3.0× LM Studio
- First token, 37k prompt
- 3.2×
- ~64s vs Ollama’s ~205s · oMLX ~69s · LM Studio ~65s
- Editing a file
- 3.1×
- 171.2 tok/s vs Ollama’s 55.6 · 3.4× oMLX · 6.7× LM Studio
Privacy and security
Your prompts stay yours.
Prompts are not retained.
Prompts and generated responses are not retained. Midium does not train on them.
Vaults stay until you delete them.
Vaults persist until you delete them.
SOC 2 is in progress.
SOC 2 Type II targeted by January 15, 2027.
Cloud API
One API name. The model underneath can change.
Instead of learning everything about every model released, simply select the model class you want and pay rates based on those tiers. We stay current. We measure the models ourselves, keep up with the latest releases, and select the best option for each model class. You do not change a thing in production. New accounts start with $5 free.
| API | Model name | Input / M | Output / M |
|---|---|---|---|
| LiteGemma 4 E2B | lite | $0.08 | $0.40 |
| BalancedGemma 4 26B | balanced | $0.18 | $0.98 |
| FrontierOrnith 1.5 122B | frontier | $0.95 | $3.73 |
| Embeddings | midium | $0.012 | $0.036 |
| System OneDecider 4B | midium | $0.032 | $0.00 |
When volume is real
Nervous about scaling usage costs?
Start on tokens like anywhere else to build as cheaply as possible. Once you move into production and costs scale with usage, switch to our dedicated capacity to keep costs flat while usage scales.
Cost vs usage
Usage-based pricing
Dedicated capacity
Usage-based pricing climbs. Dedicated capacity stays flat.
Create an account and start using the API today.
The desktop app is there when you want it on a Mac.