DeepSWE
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
claude-sonnet-5-5sonnet alias now selects it. Sonnet 5 stays selectable in every picker it already appears in.The API id is claude-sonnet-5-5 on the Claude API, Google Cloud, Microsoft Foundry, and Claude Platform on AWS; Amazon Bedrock uses anthropic.claude-sonnet-5-5. Continuum uses the bare id at the gateway and continuum/claude-sonnet-5-5 in cross-provider pickers.
Adaptive thinking is on by default. Effort levels are low, medium, high, xhigh, and max; the API default is high. The lowest thinking setting is between_tools, which works at high effort or below. Sending disabled, a thinking budget, or a non-default temperature, top_p, or top_k returns a 400. The input context is 1,000,000 tokens, maximum output is 128,000 tokens, and the knowledge cutoff is June 2026. Input is text and images; output is text.
Anthropic positions it as the best combination of speed and intelligence, and says it runs more than 30% faster per task than Sonnet 5 at the same price.
Use Sonnet 5.5 as the everyday Claude for new sessions: plan mode, coding runs, and agent work where Haiku is too small but Opus or Fable would be more than the task needs.
Choose it instead of Sonnet 5 for every new session. The old id remains selectable, so nothing already running needs to move.
No independent board lists this exact SKU as of 2026-08-19.
Closest benchmarked sibling: Claude Opus 5, 74% ±4% Pass@1 at max on the DeepSWE official board. That is a score for the sibling, not for this SKU.
Boards checked: DeepSWE official (2026-08-13), Artificial Analysis Intelligence Index, vals.ai hero index. Every card that does carry a row is on the coding leaderboard.
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
No Intelligence Index integer fetched for this identity. Chip omitted.
We did not open a vals.ai card for this identity. Chip omitted.
disabled. That returns an error; send between_tools at high effort or below, or set a lower effort, instead.computer_20251124 computer use tool is not accepted.Get Plus to use the Continuum-hosted Sonnet 5.5 rail. Not a Mac download.
Anthropic does not publish Sonnet 5.5 weights. There is no official Hugging Face repo and no Files button on this page. Get Plus buys hosted inference, not a local checkpoint.
tool_choice values (any or a named tool) return a 400. Automatic tool selection remains available.Lab card and OpenRouter listing only. No Hugging Face Files button, because there is no official repo to point at.
Fetched OpenRouter catalog on 2026-08-19. hugging_face_id was empty.
Continuum hosts the bare gateway id claude-sonnet-5-5. Native Claude sessions use the same bare id on the user’s Anthropic subscription.
The previous claude-sonnet-5 row stays in new-session menus below Sonnet 5.5. Its pricing and hosted capability metadata are unchanged.
claude-sonnet-5-5
click to select
We list first-party and OpenRouter, plus Continuum only when the live public allowlist named the id. This is not a 15-host routing table.
Official pricing per million tokens: $2 input, $10 output, $0.20 cache read, $2.50 five-minute cache write, and $4 one-hour cache write.
Cache reads are 0.1x input. There is no long-context tier: the full 1M window bills at the standard rate.
On a Claude plan? The plan picker models the weekly caps.
| Claim | Source | As of |
|---|---|---|
| API ids, context, output ceiling, adaptive thinking, default effort, knowledge cutoff, modalities, breaking changes vs Sonnet 5 | Anthropic Sonnet 5.5 overview | 2026-09-28 |
| $2 input / $10 output / $2.50 5m write / $4 1h write / $0.20 cache read; no long-context tier | Anthropic pricing | 2026-09-28 |
| No base OpenRouter slug on the 2026-08-19 catalog; Hugging Face unlisted | OpenRouter /api/v1/models | 2026-08-19 |
Same official DeepSWE harness (mini-swe-agent, public 113 tasks), fetched 2026-08-19. Different effort labels are the lab's own setting on that board, shown here rather than normalized.
Continue through Anthropic's model family with Claude Fable 5.
It takes the sonnet alias and the top Sonnet slot in every picker, but Sonnet 5 stays selectable everywhere it already is. Nothing already running needs to move.
Up-front thinking turns off with thinking type between_tools at high effort or below. Sending disabled returns an error, as does between_tools at xhigh or max effort.
No. Forced tool choice returns an error. Leave selection on automatic. Continuum drops forced choices before forwarding the request.
Lab posts pick a harness, an effort, a split, and sometimes a private eval set. DeepSWE publishes the official mini-swe-agent row with cost, tokens, and steps. Scale SWE-bench Pro only counts when the public shared-harness board has a row we can fetch. A higher lab number is usually a different test, not a better one.
Verified against the live public allowlist on 2026-08-19. Base URL is https://continuumcode.ai/v1. Keys are cont_sk_ from Settings, Account, Inference API. Personal keys need Plus or above.
curl https://continuumcode.ai/v1/chat/completions \
-H "Authorization: Bearer $CONTINUUM_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"claude-sonnet-5-5","messages":[{"role":"user","content":"Review this diff."}]}'