DeepSWE
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
gpt-5.6-terracodex -m gpt-5.6-terra. See Codex models.Closed. Same family as Luna and Sol. OpenRouter catalog 19 Aug 2026: 1,047,576 in / 128,000 out; created unix 1783590857 (9 Jul 2026).
Official OpenAI pricing.md Standard ≤272k is the $2 / $12 pair. Long-context >272k is $4 / $18. Do not invoice the long-context pair as the everyday list.
Use Terra when Luna is too small and Sol is more than you need. AA Intelligence Index 57; vals hero 56.53% ±1.24. We do not invent a DeepSWE percent: that board has no Terra row in the 13 Aug family notes this cluster already cites.
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
vals.ai Accuracy Rankings hero for Terra, fetched 19 Aug 2026: 56.53% ±1.24, latency 11 min 40 s. That is the vals hero, not DeepSWE, and not SWE-bench Verified as a Continuum chip.
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.
Copied from AA or vals HTML we opened. 0.0% placeholder bars are omitted. Do not average these into the hero tiles, and do not invent CI, $/task, or steps for Artificial Analysis.
| Bench | Printed | Note | Source | URL | As of |
|---|---|---|---|---|---|
| AA throughput | 119 tok/s | Printed Speed row. Not a coding score. | Artificial Analysis GPT-5.6 Terra | source | 2026-08-19 |
| AA Index eval cost | $1,389.66 | Printed Index-eval total. Not DeepSWE $/task. | Artificial Analysis GPT-5.6 Terra | source | 2026-08-19 |
| vals hero extras | 11 min 40 s latency | Printed latency next to Accuracy 56.53% ±1.24. | vals.ai Accuracy Rankings 2026-08-19 | source | 2026-08-19 |
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.
OpenAI does not publish Terra weights. No HF Files button. 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 gpt-5.6-terra on the 19 Aug 2026 bundled allowlist, with Luna and Sol.
gpt-5.6-terra
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.
No DeepSWE $/task: that board has no Terra row.
Official OpenAI pricing.md, 19 Aug 2026: Standard $2 / $12; cache $0.20 / $2.50 write; >272k $4 / $18. OpenRouter reprints the standard pair plus $0.01/request search.
Artificial Analysis Intelligence Index 57 at max, fetched 19 Aug 2026. Same page: 119 output tokens/s; $1,389.66 to run the Index. The tile is the published integer 57. Do not invoice at the Index-eval total.
| Claim | Source | As of |
|---|---|---|
| Official Terra Standard $2 / $0.20 cached / $2.50 write / $12; >272k $4 / $0.40 / $5 / $18 | OpenAI pricing.md | 2026-08-19 |
| OR catalog $2 / $12, cache r/w $0.20 / $2.50, search $0.01, 1,047,576 / 128,000, created 1783590857 | OpenRouter catalog | 2026-08-19 |
| AA Index 57 (max); 119 tok/s; $1,389.66 Index eval | Artificial Analysis GPT-5.6 Terra | 2026-08-19 |
| vals Accuracy 56.53% ±1.24; 11 min 40 s | vals.ai Accuracy Rankings 2026-08-19 | 2026-08-19 |
| Hosted id gpt-5.6-terra | Continuum HostedKeyProvider bundled catalog | 2026-08-19 |
| 943B week tokens (+12% WoW) | OpenRouter rankings This Week | 2026-08-19 |
| OpenRouter id openai/gpt-5.6-terra, context 1,047,576 | OpenRouter /api/v1/models | 2026-08-19 |
No other card in this cluster shares a published DeepSWE official row we can put next to this one. We will not compare on vendor-blog numbers.
Continue through OpenAI's model family with gpt-oss-120b.
The official mini-swe-agent board notes this cluster already cites name Fable, Opus, V4 Pro, and Gemini 3.6 Flash, not Terra. vals 56.53% is a different harness.
Yes. The bundled HostedKeyProvider allowlist names gpt-5.6-terra with Luna and Sol.
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":"gpt-5.6-terra","messages":[{"role":"user","content":"Review this diff."}]}'