Skip to content
Model Pareto

Hermes 4 - Llama-3.1 70B

Nous ResearchOpen weightsartificialanalysis.ai ↗
Compare with nearest rivalsCompare with…
Input
—
Output
—
Speed
~67tok/sest.
Context
128Ktokens

Long-tail import from Artificial Analysis (2026-09-24); AA variant: Hermes 4 - Llama-3.1 70B (Non-reasoning). Price = AA median host price.

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs. No API sells this model, so this is its price on the chart.

Hardware
1× B200 180GB
FP8 weights
Throughput
~2,313 tok/s
many requests batched
Price
$0.35–3.12 /M tok
busy → light use
On your own machine
Mac Studio M5 Max 128GB (int4, ~12 tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • FP8 weights (70.6B params) + 25% KV-cache headroom ≈ 88 GB
  • 1× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
  • ~2,313 output tok/s aggregate at batch 66 (bandwidth-bound)
  • Blended 3:1 input:output; prefill ~15,058 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Writing & chat

MMMLUEstimated
~last of 5 (estimated)below measured range (<81.3%)
  1. #1Gemini 3.1 Pro92.6%
  2. #2Gemma 4 31B88.4%
  3. #4gpt-oss-120b81.3%
  4. ~#5Hermes 4 - Llama-3.1 70BEstimated<81.3%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Estimated
~#49 of 54 (estimated)~1295 Elo
  1. #1Claude Opus 5.51501
  2. #2Claude Fable 5.1Few votes1496
  3. #48gpt-oss-20b1299
  4. ~#49Hermes 4 - Llama-3.1 70BEstimated1295
  5. #49Olmo 3.1 32B Instruct1294
No published score — estimated from related benchmarks
LMArena Text (Elo)Estimated
~#49 of 54 (estimated)~1322 Elo
  1. #1Claude Opus 5.51509
  2. #2Claude Fable 5.11501
  3. #48Molmo2-8BFew votes1322
  4. ~#49Hermes 4 - Llama-3.1 70BEstimated1322
  5. #49Ring-flash-2.01321
No published score — estimated from related benchmarks

Hard reasoning

AIME (latest)Estimated
~#18 of 25 (estimated)~32.4%
  1. #1Inkling97.1%
  2. #2Muse Glimmer94.7%
  3. #17Phi-4-mini-flash-reasoning33.6%
  4. ~#18Hermes 4 - Llama-3.1 70BEstimated32.4%
  5. #18Ministral 3 8B31.7%
No published score — estimated from related benchmarks

Agents

OSWorld-VerifiedEstimated
~#11 of 12 (estimated)~62.1%
  1. #1Kimi K384.8%
  2. #2Qwen3.8 27B84.3%
  3. #10Muse Glimmer65.9%
  4. ~#11Hermes 4 - Llama-3.1 70BEstimated62.1%
  5. #11Claude Haiku 4.550.7%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#3 of 5 (estimated)~53.8%
  1. #1GLM-5.3 Flash59.1%
  2. #2Kimi K358.3%
  3. ~#3Hermes 4 - Llama-3.1 70BEstimated53.8%
  4. #3GPT-5.6 Terra50.2%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#17 of 18 (estimated)~31.2%
  1. #1Atria Dawn Preview92.5%
  2. #2GPT-6 Astra91.5%
  3. #16Nemotron 3 Ultra44.4%
  4. ~#17Hermes 4 - Llama-3.1 70BEstimated31.2%
  5. #17Mistral Small 421.3%
No published score — estimated from related benchmarks
GDPval-AA (Elo)Estimated
~#53 of 87 (estimated)~728 Elo
  1. #1Claude Opus 5.51846
  2. #2Claude Fable 5.11735
  3. #52Ring-2.6-1T734
  4. ~#53Hermes 4 - Llama-3.1 70BEstimated728
  5. #53KAT Coder Pro V2720
No published score — estimated from related benchmarks
τ-Bench Banking (AA)Estimated
~#89 of 89 (estimated)~3.3%
  1. #1Muse Spark 1.350.5%
  2. #2GLM-5.350.3%
  3. #88Llama 4 Scout3.3%
  4. ~#89Hermes 4 - Llama-3.1 70BEstimated3.3%
No published score — estimated from related benchmarks