Anthropic · in Borg

Claude Opus 4.6

Claude Opus 4.6 is an earlier flagship Opus release, built for difficult coding, long agent runs, and knowledge work.

Reasoning1,000,000 tokenstext, image

1,000,000 tokensContext window

—Median time to first token (gateway proxy)

35.5Output tokens per second (whole turn)

$5 / $25Vendor API price in / out · per 1M

Included in Borg

Download Borg Borg for Business

At a glance

Specification	Value
Provider	Anthropic
Family	Claude
Release date	2026-02-05 Vendor source
Context window	1,000,000 tokens Vendor source
Input modalities	text, image
Output modalities	text
Reasoning	Yes
Open weights	No
Knowledge cutoff	2025-05 Vendor source
Borg lane ids	`antigravity/claude-opus-4-6-thinking`
Versions in Borg	Claude Opus 4.6 · Claude Opus 5.5
Median tokens per second (end to end) on Borg	35.5 output tokens per second

Not published by the vendor: thinking modes, fusion roles.

What Claude Opus 4.6 is good for

Claude Opus 4.6 is an earlier flagship Opus release, built for difficult coding, long agent runs, and knowledge work. It improved on Opus 4.5 in planning, sustained tool use, and handling large bodies of material; adaptive thinking gives it more room to deliberate on difficult work. Anthropic now classifies it as a legacy model and recommends the newer Opus 5.5, which has a lower published base price. Opus 4.6 remains useful when an established workflow depends on this generation’s behavior or when comparing results against a newer Opus run. Its large context and output limits suit repository-scale analysis and substantial reports, rather than cheap routine classification. Borg includes Opus 4.6 in the subscription; Borg Fusion can route suitable work to it, and its lane can be selected directly for continuity.

maintaining workflows calibrated to Opus 4.6
agentic coding with long tool sequences
large-repository analysis and refactoring
research reports spanning many documents

How Borg uses it

Claude Opus 4.6 is available as a subscription catalog lane in Borg. The catalog does not publish which Fusion profiles route to this model, so no routing role is claimed here. You can pick it directly by setting the OpenAI-compatible gateway model field to a listed lane id. Check the current catalog for availability before building a workflow around a lane.

model: "antigravity/claude-opus-4-6-thinking"

Measured on Borg

Metric	Median	95th percentile	Samples
Time to first token (gateway proxy)	—	—	0
Tokens per second (end to end)	35.5 output tokens per second	62.7 output tokens per second	153

Output tokens per second (whole turn)

Higher is fasterFleet median 31.9 tokens/s

Claude Opus 4.635.5 tokens/s

Time to first token (gateway proxy): time until the first streamed response event reaches the gateway; our closest available measure of time to first token
Tokens per second (end to end): tokens per second, end to end: provider-reported output tokens divided by the whole turn's elapsed time, including queueing and reasoning; not a streaming per-token rate

Completion speed covers the whole turn including queueing and reasoning time; it is not a streaming tokens-per-second figure.

Measured on Borg: SearchAtlas AI Gateway production request logs. 30-day window ending 2026-10-08; 191 requests. Minimum 100 samples per published request metric.

Benchmarks

Benchmark	Vendor-reported score	Source
Terminal-Bench 2.0	65.4%	Vendor source
SWE-bench Verified	80.8%	Vendor source
OSWorld	72.7%	Vendor source
BrowseComp	84.0%	Vendor source
Humanity’s Last Exam (without tools)	40.0%	Vendor source

Pricing

Included in Borg subscription. Vendor API list price: $5 input / $25 output per 1M tokens. Vendor price source.

Compare subscription and API pricing · Cost Savings Calculator

FAQ

What is Claude Opus 4.6?

Claude Opus 4.6 is listed in Borg's subscription catalog as antigravity/claude-opus-4-6-thinking. Claude Opus 4.6 is an earlier flagship Opus release, built for difficult coding, long agent runs, and knowledge work.

Is Claude Opus 4.6 in Borg?

Yes. Claude Opus 4.6 is listed in Borg's subscription catalog under the lane antigravity/claude-opus-4-6-thinking. You can choose the lane directly, or explore the available Fusion profiles. Availability and model details can change; check the current catalog before deployment.

What context window does Claude Opus 4.6 have?

The published context window is 1,000,000 tokens. Consult the cited sources for details and current limits. Vendor source

Is Claude Opus 4.6 good for coding?

Vendor-published evaluations appear in the benchmarks section; their settings may differ from your tasks. Its listed uses include maintaining workflows calibrated to Opus 4.6, agentic coding with long tool sequences, large-repository analysis and refactoring, research reports spanning many documents. Try the lane for your own tasks, and compare the output with your requirements before choosing a default.

How fast is Claude Opus 4.6 on Borg?