DeepSWE
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
Independent lab flagship. We card the 405B. AA title is “Hermes 4 - Llama-3.1 405B (Non-reasoning)”; Intelligence Index 9. No official DeepSWE or vals card. Hub base_model is Meta-Llama-3.1-405B.
Independent lab, not a frontier closed API. Llama 3 license on HF. 131k context. OpenRouter catalog created unix 1756235463 (26 Aug 2025). Max output unpublished on that catalog row.
HF config.json (fetched 19 Aug 2026): LlamaForCausalLM, 126 layers, hidden 16384, 128 attention heads, 131,072 positions. Hub base_model is meta-llama/Meta-Llama-3.1-405B.
Use Hermes 4 when you want Nous’s open 405B. AA Intelligence Index 9 is the Non-reasoning tile. No official DeepSWE or vals card.
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
We did not open a vals.ai card for this identity. Chip omitted.
vals.ai /models/nousresearch_hermes-4-405b returned 404 on 19 Aug 2026. Chip omitted.
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 | $1.00 / $3.00 per 1M | Printed AA pricing sentence on the Non-reasoning tile. Matches OpenRouter. | Artificial Analysis Hermes 4 405B | source | 2026-08-19 |
| AA throughput | 28.1 tok/s | Printed Speed row. Not a coding score. | Artificial Analysis Hermes 4 405B | 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.
Community quants: Community quants are community.
This is the model repo, not Download for Mac. Downloads and likes from the Hugging Face API on 2026-08-19.
usedStorage and the sum of .safetensors sizes can differ (LFS pointers, non-shard files, duplicate copies). Cited separately. Source: recursive tree API, 2026-08-19.
| Path | Bytes |
|---|---|
model-00001-of-00191.safetensors | 4,806,672,880 |
model-00002-of-00191.safetensors | 4,026,532,224 |
model-00003-of-00191.safetensors | 4,630,578,112 |
model-00004-of-00191.safetensors | 4,630,578,112 |
model-00005-of-00191.safetensors | 3,489,661,192 |
model-00006-of-00191.safetensors | 4,630,578,112 |
model-00007-of-00191.safetensors | 4,630,578,112 |
model-00008-of-00191.safetensors | 3,489,661,192 |
model-00009-of-00191.safetensors | 4,630,578,112 |
model-00010-of-00191.safetensors | 4,630,578,112 |
model-00011-of-00191.safetensors | 3,489,661,192 |
model-00012-of-00191.safetensors | 4,630,578,112 |
Showing the first 12 of 191 safetensor shards. The tree also holds 9 non-shard files, none of them listed here. The snapshot was capped at 12 shards when the tree API was read on 2026-08-19, so the remaining 179 are on the Files tab, not on this page. Totals above are the full tree API sum, not this slice. Byte sizes are Hub tree API size fields, never invented.
Not listed as a Continuum hosted id. OpenRouter nousresearch/hermes-4-405b.
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.
OpenRouter catalog 19 Aug 2026: $1 / $3 per million, 131,072 in, max out unpublished, created 1756235463. No cache fields. No DeepSWE $/task.
Artificial Analysis page title is “Hermes 4 - Llama-3.1 405B (Non-reasoning)”. That page prints $1.00 / $3.00 and 28.1 tok/s. Cite the title; do not invent a reasoning-mode Index.
Artificial Analysis title is “Hermes 4 - Llama-3.1 405B (Non-reasoning)”. Intelligence Index 9, fetched 19 Aug 2026. Same page prints $1.00 / $3.00 and 28.1 tok/s. The AA tile is the published integer 9. AA did not print a DeepSWE-style CI, step count, or Index-eval cost we will copy onto the chip.
| Claim | Source | As of |
|---|---|---|
| OR catalog $1 / $3, 131,072 in, max out unpublished, created 1756235463, hugging_face_id NousResearch/Hermes-4-405B | OpenRouter catalog | 2026-08-19 |
| HF safetensors.total 405,853,388,800; usedStorage 811,724,126,627; 191 shards; likes 94; downloads 736; created 2025-08-06T02:51:17Z; arXiv:2508.18255; Llama 3 license; base_model Meta-Llama-3.1-405B | Hugging Face API NousResearch/Hermes-4-405B | 2026-08-19 |
| AA title Non-reasoning; Index 9; $1.00 / $3.00; 28.1 tok/s | Artificial Analysis Hermes 4 405B | 2026-08-19 |
| vals 404 | vals.ai Hermes 4 | 2026-08-19 |
| OpenRouter id nousresearch/hermes-4-405b, context 131,072 | OpenRouter /api/v1/models | 2026-08-19 |
| HF downloads 736 | Hugging Face API NousResearch/Hermes-4-405B | 2026-08-19 |
| 191 safetensor shards; shard bytes 811,706,916,800; tree files 200 | Hugging Face tree API NousResearch/Hermes-4-405B | 2026-08-19 |
| Spaces API returned 13; official collection Hermes 4 Collection (4 items, 119 upvotes) | Hugging Face Spaces / collections API | 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 Nous Research's model family with Hermes 4 70B.
The official DeepSWE board updated 13 Aug 2026 did not list Hermes 4 405B. Omitting is the rule.
No. AA’s own title is the Non-reasoning tile. We cite that title next to the integer.
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.
No Continuum hosted id is verified for this model, so this page does not invent a curl target. Use the first-party API or the OpenRouter slug nousresearch/hermes-4-405b.