Claude Opus 5.5: what changed and how to use it on Kumo

On September 22, 2026 Anthropic released Claude Opus 5.5 — a model that holds Claude Fable 5.1's level on most real work while costing 2.5 times less than it and 20% less than the previous Opus 5. It is already available on Kumo: the same API, one balance topped up by card or crypto, and no weekly quotas. And it is the original Opus 5.5, not a cheaper model wearing its name — below are two checks you can run yourself to confirm it.

What Claude Opus 5.5 is

What changed since Opus 5

What exactly got cheaper

What Claude Opus 5.5 costs

Opus 5.5 is already available on Kumo

export KUMO_API_KEY="kumo_sk_…"

# OpenAI-compatible clients
export OPENAI_BASE_URL="https://api.kumorouter.com/v1"
export OPENAI_API_KEY="$KUMO_API_KEY"

# clients speaking the Anthropic Messages protocol
export ANTHROPIC_BASE_URL="https://api.kumorouter.com"
export ANTHROPIC_AUTH_TOKEN="$KUMO_API_KEY"

The model is the original — and you can prove it

Check 1. The catalog and the model echo in the response

# the model must appear in the catalog under its real identifier
curl -s https://api.kumorouter.com/v1/models \
  -H "Authorization: Bearer $KUMO_API_KEY" | grep -o '"anthropic/claude-opus-5-5"'

# the model field in the response must repeat exactly the name you sent
curl -s https://api.kumorouter.com/v1/chat/completions \
  -H "Authorization: Bearer $KUMO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"anthropic/claude-opus-5-5","messages":[{"role":"user","content":"ping"}],"max_tokens":16}' \
  | python3 -c "import sys,json; d=json.load(sys.stdin); print(d['model'], d['usage'])"

Check 2. The 1M window and a needle in the haystack

python3 - <<'PY'
import os, httpx

needle = " The pass phrase of the day is kumo-opus-55-million. "
filler = "This is filler text whose only job is to fill the context window. " * 20000
question = " Question: state the pass phrase of the day on one line, with no explanation."
prompt = filler + needle + filler + question

r = httpx.post(
    "https://api.kumorouter.com/v1/chat/completions",
    headers={"Authorization": "Bearer " + os.environ["KUMO_API_KEY"]},
    json={
        "model": "anthropic/claude-opus-5-5",
        "messages": [{"role": "user", "content": prompt}],
        "max_tokens": 64,
    },
    timeout=900,
)
body = r.json()
print(body["usage"]["prompt_tokens"], "input tokens")
print(body["choices"][0]["message"]["content"])
PY

What each check proves, and what it does not

How to connect Opus 5.5 in five minutes

1. Sign up in the console and top up the balance — by card, through SBP or in crypto.
2. Create a key and give it a spending cap right away: one key per project saves you the argument about where the money went.
3. Point your client at the gateway — https://api.kumorouter.com/v1 for OpenAI-compatible tools, the bare origin for clients on the Anthropic protocol.
4. Set the model to anthropic/claude-opus-5-5 and send the first request.
5. Run checks 1 and 2 from this article before you start spending in earnest.

curl -s https://api.kumorouter.com/v1/chat/completions \
  -H "Authorization: Bearer $KUMO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-opus-5-5",
    "messages": [
      {"role": "user", "content": "Find every place in this repository where an environment variable is read and put them in a table."}
    ],
    "max_tokens": 2048
  }'

Frequently asked questions

How does Opus 5.5 differ from Fable 5.1, and which should I pick?

Fable 5.1 remains the top model for the heaviest reasoning and the longest agentic sessions. Opus 5.5 covers most of the same work for $4 and $20 against $10 and $50, and answers faster. The sensible order is to start on Opus 5.5 and move up to Fable 5.1 where your own evals show it falling short.

Is this really the same Opus 5.5 as Anthropic's own?

Yes: the route leads to the original model, not to a stand-in. You can confirm it two ways from the section above — the catalog and the identifier echo, and the million-token window with a needle in it. You run both yourself, with your own key, and the console's request log shows what each request was billed as.

Why is it cheaper on Kumo than at the vendor?

The catalogue discount is 40% off the vendor's list on ordinary input and output tokens, and on Opus 5.5 at launch it goes deeper: 45%, which is $2.2 and $11 instead of $4 and $20. A package ladder applies on top for larger volumes. Compare the same model version and the same kind of token: uncached input, cache reads and cache writes are billed at different rates.

Do I need a subscription, and are there weekly quotas?

No. Billing is pay-as-you-go from one balance: top it up, work, top it up again. There are no weekly or daily subscription quotas here, and the spending limits are the ones you set yourself on each key.

When do Sonnet 5.5 and Haiku 5.5 arrive?

Anthropic promises them in the coming weeks with similar gains. We connect new vendor models as they ship — watch the models page and the blog.

Sources

• Anthropic's announcement: Introducing Claude Opus 5.5 — benchmarks, pricing and safety data.
• Vendor documentation: Claude models overview — identifiers, context window, knowledge cutoffs and retirement dates.
• Anthropic pricing: Pricing — cache rates, the Batch API and per-cloud prices.
• Kumo pricing: the models page — the rates in force for every model in the catalog.