Model comparison
Claude Opus 5.5 vs GPT-5.6 Luna
Claude
GPTCompare published specifications, vendor API list prices and measured Borg performance. Both models are included in Borg.
Claude Opus 5.5
GPT-5.6 Luna
Included in Borg
At a glance
| Specification | Claude Opus 5.5 | GPT-5.6 Luna |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Release | 2026-09-22 Vendor source | 2026-07-09 Vendor source |
| Context | 1,000,000 tokens Vendor source | 1,050,000 tokens Vendor source |
| Modalities | Input: text, image; output: text | Input: text, image; output: text |
| Reasoning / thinking modes | Yes / Not published | Yes / Not published |
| Open weights | No | No |
| Vendor API list price / 1M tokens | $4 input / $20 output | $0.2 input / $1.2 output Standard global rates for prompts up to 272,000 input tokens. Above 272,000 input tokens, input and cached-input rates are 2× and output rates are 1.5× for the full request. |
| Borg | Included in Borg subscription | Included in Borg subscription |
AnthropicMeasured on Borg
Metrics are not directly comparable to vendor benchmark claims. Models served different requests; this is observed gateway traffic, not a controlled head-to-head test.
Time to first token (gateway proxy)
Output tokens per second (whole turn)
| Metric | Claude Opus 5.5 | GPT-5.6 Luna |
|---|---|---|
| Time to first token (gateway proxy) p50 | 2946 ms (75410 samples) | 3998 ms (22631 samples) |
| Tokens per second (end to end) p50 | 60 output tokens per second (203387 samples) | 17.2 output tokens per second (605485 samples) |
| Request success rate | Not published | Not published |
| Coding bench pass rate | Not published | 73% (15 samples) |
Measurement window, samples and methodology
The fleet reference is the median of published model medians, not a request-weighted result. Prompt lengths, reasoning settings and traffic differ between models.
Claude Opus 5.5: Measured on Borg: SearchAtlas AI Gateway production request logs. 30-day window ending 2026-10-08; 220887 requests. Minimum 100 samples per published request metric.
GPT-5.6 Luna: Measured on Borg: SearchAtlas AI Gateway production request logs. 30-day window ending 2026-10-08; 752719 requests. Minimum 100 samples per published request metric.
Completion speed covers the whole turn including queueing and reasoning time; it is not a streaming tokens-per-second figure.
- 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
How to choose
Claude Opus 5.5 is listed for long-running agentic coding across repositories, whereas GPT-5.6 Luna is listed for high-volume classification. Choosing between Claude Opus 5.5 and GPT-5.6 Luna requires checking published specifications against observed Borg traffic for your workflow.
Key differences
| Spec | Claude Opus 5.5 | GPT-5.6 Luna |
|---|---|---|
| Context window | 1,000,000 tokens Vendor source | 1,050,000 tokens Vendor source Larger context |
| Reasoning modes | Yes / Not published | Yes / Not published |
| Release date | 2026-09-22 Vendor source | 2026-07-09 Vendor source |
| API price input / 1M | 4 | 0.2 Standard global rates for prompts up to 272,000 input tokens. Above 272,000 input tokens, input and cached-input rates are 2× and output rates are 1.5× for the full request. Lower listed rate |
| API price output / 1M | 20 | 1.2 Standard global rates for prompts up to 272,000 input tokens. Above 272,000 input tokens, input and cached-input rates are 2× and output rates are 1.5× for the full request. Lower listed rate |
| Time to first token (gateway proxy) | 2946 ms (75410 samples) Lower median | 3998 ms (22631 samples) |
| Output tokens per second (whole turn) | 60 output tokens/s (whole turn) (203387 samples) Higher measured rate | 17.2 output tokens/s (whole turn) (605485 samples) |
| Request success | Not published | Not published |
| Coding bench | Not published | 73% (15 samples) |
FAQ
How do Claude Opus 5.5 and GPT-5.6 Luna compare for coding?
Claude Opus 5.5 has a Borg coding benchmark of not published; GPT-5.6 Luna has 73.33% pass rate across 15 samples and 3 tasks. The Claude Opus 5.5–GPT-5.6 Luna coding comparison cannot establish an overall winner from this small test alone.
How quickly do Claude Opus 5.5 and GPT-5.6 Luna start responding on Borg?
Borg median first response for Claude Opus 5.5 is 2,946 ms; for GPT-5.6 Luna it is 3,998 ms. The Claude Opus 5.5–GPT-5.6 Luna gateway medians are observations, not guaranteed latency for individual requests.
What are the API prices for Claude Opus 5.5 and GPT-5.6 Luna?
The listed API rate for Claude Opus 5.5 is $4 input / $20 output per 1M tokens; for GPT-5.6 Luna it is $0.2 input / $1.2 output per 1M tokens. Neither Claude Opus 5.5 nor GPT-5.6 Luna has its Borg subscription charge represented by those provider API rates. Claude Opus 5.5 vendor source · GPT-5.6 Luna vendor source