DeepSWE
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
glm-5.3OpenRouter hugging_face_id is empty for 5.3. glm-latest aliases here as of 18 Aug 2026. GLM 5.2 publishes zai-org/GLM-5.2. This card will not put 5.2 downloads or the 5.2 DeepSWE row on 5.3.
Use 5.3 when you want the current Z.ai id Continuum actually serves. AA Intelligence Index 60; vals hero index 71.48%. DeepSWE lists GLM 5.2 only, so that chip is omitted here.
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
vals.ai GLM-5.3 card, fetched 19 Aug 2026: Index 71.48%, printed $1.40 / $4.40, 13 min 25 s latency. Those extras are not DeepSWE. Accuracy Rankings bars on that HTML printed 0.0% placeholders: omitted. No Updates prose with named non-zero benches was present.
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 list pair | $1.40 / $4.40 per 1M | Printed AA pricing sentence for GLM-5.3 (max). Matches official Z.ai and OpenRouter. | Artificial Analysis GLM-5.3 | source | 2026-08-19 |
| AA throughput | 74.0 tok/s | Printed Speed row. Not a coding score. | Artificial Analysis GLM-5.3 | source | 2026-08-19 |
| AA Index eval cost | $1238.50 | Printed “it cost $1238.50 to evaluate GLM-5.3 (max) on the Intelligence Index.” | Artificial Analysis GLM-5.3 | source | 2026-08-19 |
| vals Index extras | 13 min 25 s latency | Printed Latency on the vals hero. Hero Accuracy is the 71.48% chip, not this footnote. | vals.ai GLM-5.3 | 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.
No official HF id for GLM 5.3 on the OpenRouter catalog we fetched. No fake weights button. 5.2 weights live on the GLM 5.2 card.
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 glm-5.3. Rate card $1.40 / $4.40 per million (19 Aug 2026).
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 for 5.3: that board lists 5.2 only.
Official Z.ai + OpenRouter + Continuum rate card, 19 Aug 2026: $1.40 / $4.40 per million, cache read $0.26 /M. Official page also prints Web Search $0.01/use and storage Limited-time Free. We will not invent a storage dollar after that label.
Artificial Analysis title is “GLM-5.3 (max)”. Intelligence Index 60, fetched 19 Aug 2026. Same page prints $1.40 / $4.40, 74.0 tok/s, and “it cost $1238.50 to evaluate GLM-5.3 (max) on the Intelligence Index.” The AA tile is the published integer 60. Do not invent CI, steps, or DeepSWE $/task for 5.3.
| Claim | Source | As of |
|---|---|---|
| Official GLM-5.3 $1.4 input / $0.26 cached / storage Limited-time Free / $4.4 output per 1M; Web Search $0.01/use | Z.ai pricing.md | 2026-08-19 |
| OR catalog $1.40 / $4.40, cache read $0.26/M, 1,048,576 / 131,072, created 1787086655, hugging_face_id empty | OpenRouter catalog | 2026-08-19 |
| Hosted id glm-5.3 | GET /v1/chat/hosted/models/public | 2026-08-19 |
| 4.34T week tokens belong to glm-5.2 | OpenRouter rankings This Week | 2026-08-19 |
| AA title GLM-5.3 (max); Index 60; $1.40 / $4.40; 74.0 tok/s; $1238.50 to evaluate Index | Artificial Analysis GLM-5.3 | 2026-08-19 |
| vals Index 71.48%; 13 min 25 s; $1.40 / $4.40 | vals.ai GLM-5.3 | 2026-08-19 |
| OpenRouter id z-ai/glm-5.3, context 1,048,576 | 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 Z.ai's model family with GLM 5.2.
That is GLM 5.2 on DeepSWE official. Pasting it onto 5.3 would be a fabricated identity.
AA’s own title is “GLM-5.3 (max)”. The integer is that tile. We do not invent a default-effort Index.
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":"glm-5.3","messages":[{"role":"user","content":"Review this diff."}]}'