DeepSWEmini-swe-agent
74%±4%
- bench
- DeepSWE
- version
- v1.1
- split
- public 113 tasks
- harness
- mini-swe-agent
- effort
- max
- n
- 113
- metric
- Pass@1
- value
- 74%
- ci
- ±4%
- $/task
- $11.84
- tokens
- 118k out
- steps
- 99
claude-opus-5Same 1M / 128k window as Fable 5 and Sonnet 5 on the OpenRouter catalog. Effort ladder low through max, default high.
Use Opus when Sonnet already failed, or when the decision is architectural. DeepSWE official puts Opus at 74% Pass@1 ±4% at max, $11.84 per task, 118k output, 99 steps: the top row on that board the day we fetched it.
Ranked view: Best coding models: the independent leaderboard puts this row and every other card that carries an independent coding score on one board, with the confidence intervals left visible.
Plus is $25/mo with $25 weekly hosted usage. The Mac app stays free with your own keys. Get Plus is not Download for Mac.
Anthropic does not publish Opus 5 weights. No official HF repo. Get Plus is hosted inference.
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 claude-opus-5 on the live public allowlist (19 Aug 2026). First-party Claude Code on your Anthropic login is unchanged.
claude-opus-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.
DeepSWE official: $11.84 per task at max (118k out, 99 steps). That is the hero economics number.
Official Anthropic + OpenRouter catalog, 19 Aug 2026: $5 / $25 per million. Cache read $0.50 /M. Cache write $6.25 /M (1-hour write $10 /M). Web search $0.01 per call on the catalog. Sonnet 5 is $2 / $10. Fable 5 is $10 / $50.
On a Claude plan? The plan picker models the weekly caps.
Artificial Analysis footnote, 19 Aug 2026: about $2.34 per AA task, 56.9 tok/s. The AA tile is Intelligence Index 63. AA did not publish a DeepSWE-style CI or step count we will copy onto the chip. Do not average AA cost with DeepSWE $11.84/task.
| Claim | Source | As of |
|---|---|---|
| DeepSWE 74% ±4% at max | DeepSWE official board | 2026-08-13 |
| Official Opus 5 $5 / $6.25 5m write / $10 1h write / $0.50 hits / $25 out | Anthropic pricing.md | 2026-08-19 |
| OR catalog $5 / $25, cache read $0.50, write $6.25, 1h write $10, search $0.01 | OpenRouter catalog | 2026-08-19 |
| Hosted id claude-opus-5 | GET /v1/chat/hosted/models/public | 2026-08-19 |
| 2.68T week tokens (+89% WoW) | OpenRouter rankings This Week | 2026-08-19 |
| OpenRouter id anthropic/claude-opus-5, context 1,000,000 | 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 Sonnet 5.
On DeepSWE official they are inside each other’s confidence interval (74% ±4 vs 73% ±3) and Opus costs more per task ($11.84 vs $8.39). Pick on product fit and which subscription you already pay for, not on a one-point gap.
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-opus-5","messages":[{"role":"user","content":"Review this diff."}]}'