n

Hermes 4 - Llama-3.1 70B

Aug 2025
nHermes 4 - Llama-3.1 70B
Compare in
on
—per task?
$2.88 / $2.88per 1M in / out?
—output speed?
128Kcontext

Rankingin non-thinking mode?

Strong at

C
Tool calling#183 of 242Reasoning#322 of 334Science#259 of 312Long context#268 of 273Knowledge#202 of 318Instructions#214 of 245

Alternatives?

ModelScoreCostPer taskWhat changes
CheaperMuse Spark 1.3max×7.2-31%$1.60Only auto tool_choice is supported+5BetterGPT-6 Lunalow×3.2-93%$0.005Function calling only with none · Chat Completions+3OpenQwen3.8 27Bxhigh×5.1-67%$1.01max_tokens range · QwenCloud+4CheaperMiMo v2.6 ProthinkingOpen×7-81%$0.13Memory (4-bit): 42 GB → 615 GB+4CheaperDeepSeek V4.1 Flash 1017maxOpen×6.8-82%$0.26Memory (4-bit): 42 GB → 458 GB+5

Run it yourself

estimate
Fits on✗ 8 GB (phone, laptop)✗ 16 GB✗ 24 GB GPU✓ 48 GB GPU · compressed✓ 80 GB GPU · 8-bit
LicenceLlama 3 licencecommercial use with conditions

Benchmarks

all categories
Chat & Knowledge
Humanity's Last Exam3.6#325
AA Omniscience Index-48.6#203
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈6.7#284
AA Omniscience · Accuracy18.7#171
AA Omniscience · Hallucination Rate82.7#187
AA-LCR (long context) 1.14.7#270
GPQA Diamond49.1#262
AIME 2025 · AA archived11.3#125
Coding
Terminal-Bench Hard (version unstated)0#212
LiveCodeBench · AA archived26.9#128
Agents & Tool Use
τ²-bench · AA21.6#185
Writing & Language
IFBench (instructions)29#217

FAQ

How much does Hermes 4 - Llama-3.1 70B cost?›

$2.88 per 1M input tokens and $2.88 per 1M output tokens (cheapest host: Featherless).

Is Hermes 4 - Llama-3.1 70B free to use?›

There is no free API; the cheapest paid price is $2.88 per 1M input tokens. The weights are open, so you can run it yourself.

How good is Hermes 4 - Llama-3.1 70B, and what is it best at?›

#322 of 334 in Chat & Knowledge, #209 of 239 in Coding, #183 of 242 in Agents & Tool Use, #214 of 245 in Writing & Language.

Is Hermes 4 - Llama-3.1 70B better than Hermes 3 70B Instruct?›

On AA Intelligence Index 4.3.2, Hermes 4 - Llama-3.1 70B scores 7.9 vs 6 — higher.

How can I use Hermes 4 - Llama-3.1 70B?›

Through the API of Featherless.

What is the context window of Hermes 4 - Llama-3.1 70B?›

128K tokens.

When was Hermes 4 - Llama-3.1 70B released? Is there a newer version?›

Released 2025-08-27. It is the current version in its line.

Is Hermes 4 - Llama-3.1 70B open source?›

Yes, the weights are open (llama3 licence).

Can I run Hermes 4 - Llama-3.1 70B locally?›

The weights are open and can be self-hosted; hardware needs are not confirmed yet.

Can Hermes 4 - Llama-3.1 70B read images?›

No, it accepts text only.

What is Hermes 4 - Llama-3.1 70B?›

Hermes 4 - Llama-3.1 70B is an open-weights language model from nousresearch, released Aug 2025, the successor to Hermes 3 70B Instruct.

What are the best alternatives to Hermes 4 - Llama-3.1 70B?›

Muse Spark 1.3 (cheaper), GPT-6 Luna (better), Qwen3.8 27B (open weights).