Model overview · 2026

Reasoning model · free to try

Meet Ox Alpha.

A frontier reasoning model built for code, long-horizon agents, and a million tokens of context. Fast, capable, and free to try.

Start chatting Learn more

Research team · September 2026

At a glance

Specification

Four numbers that define it

Context window

1M

tokens held in memory at once

Max output

131K

tokens in a single response

Input modalities

3

text · image · video

Price to try

$0

no account, no card

Reasoning

Reasoning

Thinks before
it speaks.

Ox Alpha is a reasoning-first model. It plans, checks itself, and only then answers — so hard problems get worked through instead of guessed at.

  • Multi-step logic. Math, systems design, proofs, tricky edge cases — worked through, not pattern-matched.
  • Visible thinking. The reasoning streams in before the final answer lands.
  • Production-grade output. Tuned for real workloads, not party tricks.
Context window

Tokens held in memory

1,048,576

Paste an entire codebase, a 500-page spec, or a week of transcripts. Ox Alpha keeps all of it in mind while it works — no chunking, no retrieval tricks, no losing the thread.

Capabilities

Capabilities

Built for real work.

Everything a frontier model should do — and a few things most of them can't.

Flagship

Long-horizon coding

Sustained software engineering: debugging, refactoring, and reasoning across large codebases without losing context.

Vision

Sees what you see

Text, images, and video in. Screenshots, diagrams, and UI — reasoned about, not just described.

Agents

Tool calling

Native tools and tool_choice support for agents that actually do things.

Integration

Structured output

Ask for JSON and get JSON — schema-shaped responses via response_format.

Stamina

Sustained agentic work

Keeps its footing across long, multi-step tasks instead of drifting off halfway.

Access

No barrier to entry

Open the browser and start. No account, no card, no waitlist — on any device.

In action

In action

Watch it
reason.

A reasoning model shows its work. It names the failure mode before it writes the fix.

Prompt

Why does this Python function sometimes return stale data?

cache = {}
def get(k): return cache.get(k) or fetch(k)

Thinking

or treats falsy cached values (0, "", []) as misses and re-fetches. And nothing ever writes to cache — so the stale data must come from fetch itself.

Answer

def get(k):
    if k not in cache:      # explicit miss check
        cache[k] = fetch(k)   # actually populate
    return cache[k]
Benchmarks

Independent coding benchmark

80%

8 of 10 real-world coding tasks solved — ahead of every frontier model it was lined up against.

Benchmarks · comparison

Mean score, same 10 tasks

Frontier reasoning.
Top of the table.

Ox Alpha
80%
Fable 5
65%
GLM-5
62%
Grok 4
62%
GPT-5
52%

Reference models scored as passes out of 4 attempts per task; Ox Alpha as pass/fail.

Benchmarks · task detail

Ten tasks, side by side

TaskFable 5GLM-5.3GPT-5.6Grok 4.6Ox Alpha
anko-typed-variable-bindings4/44/42/41/4✓
arktype-json-schema-refs2/41/43/41/4✓
fastapi-deprecation-headers4/43/43/44/4✓
helm-unified-manifest-stream4/44/44/44/4✓
igel-persist-feature-schema3/43/40/44/4✓
katex-multicolumn-array-spans2/44/43/44/4✓
meriyah-explicit-resource-decl1/40/40/40/4✓
query-persist-restored-state2/43/41/42/4✓
scc-bounded-memory-spilling4/43/44/44/4✕
vulture-persistent-analysis-cache0/40/41/41/4✕
Mean on these 1065%62%52%62%80%

Highlighted row: the one task every reference model scored 1/4 or worse on — Ox Alpha solved it. Third-party data, small sample — directional, not definitive.

How it works

How it works

Zero to answer in seconds.

Step 01

Describe the task

Type a goal in plain English. Paste the whole file, log, or document — context is not a problem.

Step 02

Ox Alpha reasons

It thinks through the problem step by step, then streams back its answer in real time.

Step 03

Review & iterate

Refine with follow-ups. The whole conversation stays in its 1M-token memory.

Try it on

Try it on

Pick a task. Or bring your own.

Coding

Debug a gnarly traceback

Root-cause analysis from a model built for long-horizon engineering.

Reasoning

Hard math or logic

It plans before it answers — steps, not guesses.

Big context

A whole codebase in one prompt

1M tokens means the repo fits. No chunking, no RAG.

Vision

Explain a screenshot or chart

Text plus visual context, reasoned about together.

Agentic

Design a multi-step workflow

Tool calling, structured output, stamina for long tasks.

Coding

Review a pull request

Correctness, edge cases, simplifications — with reasoning.

Writing

Rough notes to a real doc

Structured, precise writing tuned for production.

Big context

Digest a 500-page PDF

Summary, quotes, action items — whole document in context.

Access & privacy

Access & privacy

Nothing to sign.
Nothing stored.

  • No account, no card, no waitlist. Open the browser and start.
  • Conversations aren't stored on our servers. No personal information is required.
  • Works everywhere. Any modern browser — phone, tablet, desktop.

Messages are processed by the upstream model API to generate responses. As with any AI service, avoid sharing sensitive data.

Free to try

Try it before the internet
figures out what it is.

No account, no card, no waitlist.

Open chat

oxalpha.com · 2026