Canonical: https://krasis.ai/benchmarks/

 Public benchmark ledger

# Explore what changes.  
_Keep the context._

Compare measured speed by GPU, PCIe link and model, inspect quality evidence, or see every supported model ordered by its persistent system RAM floor. Every row keeps the configuration details that make the number meaningful.

ExploreSpeedQualitySystem RAMGPURTX PRO 6000 · 96 GB · PCIe 4.0 ×16RTX A6000 · 48 GB · PCIe 4.0 ×16RTX 5090 · 32 GB · PCIe 4.0 ×16RTX 5090 · 32 GB · PCIe 5.0 ×16ModelAll measured modelsDeepSeek-V4-Flash-0731Qwen3-Coder-NextOrnith-1.0-397BStep-3.7-Flash

**4**measured configurations

**Speed results**

## RTX PRO 6000 · 96 GB · PCIe 4.0 ×16

Prefill is prompt processing speed. Decode is Krasis’s internal engine measurement; HTTP round trip includes the local client/server path.

Scroll sideways to inspect every column →

| Model                  | Params                     | PCIe    | Attention + KV         | Prefill        | Decode      | HTTP round trip | Peak system RAM                         | HCS                  | Min free VRAM |
| ---------------------- | -------------------------- | ------- | ---------------------- | -------------- | ----------- | --------------- | --------------------------------------- | -------------------- | ------------- |
| DeepSeek-V4-Flash-0731 | 284B main · 13B active     | 4.0 ×16 | INT4/HQQ6/Native cache | 2,620.6 tok/s  | 38.16 tok/s | 70.30 tok/s     | 152.3 GB process RAM (benchmark report) | 6660/11008 (60.5%)   | 1,189 MB      |
| Qwen3-Coder-Next       | 80B main · 3B active       | 4.0 ×16 | INT4/HQQ4/k4v4         | 11,211.1 tok/s | 91.34 tok/s | 161.82 tok/s    | 79.2 GB process RAM (benchmark report)  | 24576/24576 (100.0%) | 53,156 MB     |
| Ornith-1.0-397B        | 397B main · 17B active     | 4.0 ×16 | INT4/HQQ4/k4v4         | 2,354.5 tok/s  | 23.58 tok/s | 41.73 tok/s     | 201.8 GB process RAM (benchmark report) | 13161/30720 (42.8%)  | 1,346 MB      |
| Step-3.7-Flash         | 199.4B main · 13.9B active | 4.0 ×16 | INT4/HQQ4/k4v4         | 5,261.0 tok/s  | 55.40 tok/s | 112.82 tok/s    | 103.4 GB process RAM (benchmark report) | 11121/12096 (91.9%)  | 1,274 MB      |

How to read speed

Rows are measured configurations, not a league table. GPU VRAM and PCIe generation are separate because host-to-device bandwidth can materially change large-model decode. Prompt length, model and cache choices also affect throughput. Peak system RAM is process max RSS unless the row says otherwise.

Source snapshot

Base data mirrors the public Krasis ledgers at revision `7ab1cea`. Supplemental timing-disabled runs link their reports separately.[Speed source ↗](https://github.com/brontoguana/krasis/blob/7ab1cea59d2b5cae35356b298d0ed8bf3531b0a7/benchmarks/BENCHMARKS.md)[Quality source ↗](https://github.com/brontoguana/krasis/blob/7ab1cea59d2b5cae35356b298d0ed8bf3531b0a7/STATS-QUALITY.md)[DeepSeek PCIe 5.0 run ↗](https://github.com/brontoguana/krasis/blob/7ab1cea59d2b5cae35356b298d0ed8bf3531b0a7/benchmarks/20260814%5Fdeepseek%5Fv4%5Fhqq6%5Fnative%5Fvast%5F5090%5Fpcie5%5Fbenchmark%5Freport.log)[Ornith-397B PCIe 5.0 run ↗](https://github.com/brontoguana/krasis/blob/7ab1cea59d2b5cae35356b298d0ed8bf3531b0a7/benchmarks/20260814%5Fornith397%5Fhqq6%5Fk6v6%5Fvast%5F5090%5Fpcie5%5Fbenchmark%5Freport.log)[DeepSeek RTX A6000 run ↗](https://krasis.ai/benchmarks/20260818%5Fdeepseek%5Fv4%5Fflash%5F0731%5Fhqq6%5Fnative%5Fa6000%5Fbenchmark%5Freport.txt)