---
license: mit
task_categories:
  - text-generation
tags:
  - benchmark
  - inference
  - llm
  - nvidia
  - rtx-5090
  - llama-cpp
  - vllm
  - speed
  - quality
  - mmlu
  - gsm8k
  - humaneval
  - moe
size_categories:
  - n<1K
configs:
  - config_name: default
    data_files:
      - split: train
        path: benchmarks.csv
---

# RTX 5090 LLM Benchmarks

Speed and quality benchmarks for quantized LLMs on NVIDIA RTX 5090 32GB, measured with [llm-bench-rig](https://github.com/notwitcheer/llm-bench-rig).

## Quality Benchmarks

Generative evaluation through llama-server chat completions. Replicates standard benchmark methodology using custom evaluators — no `lm-evaluation-harness` dependency.

Results are **split by reasoning mode**: comparing a thinking-on (reasoning) model's quality against a thinking-off model is apples-to-oranges, so the two groups are ranked separately. `q_avg` is the mean of the five tasks.

### Thinking OFF (non-reasoning · direct answer)

| Model | Params | Quant | MMLU | ARC-C | HellaSwag | GSM8K | HumanEval | q_avg |
|-------|-------:|-------|-----:|------:|----------:|------:|----------:|------:|
| Gemma 4 31B-it | 30.70B | Q6_K | 87.8 | 97.6 | 92.0 | 97.5 | 96.3 | **94.2** |
| Qwopus3.6-27B-Coder | 27.32B | Q5_K_M | 87.5 | 96.8 | 95.2 | 97.5 | 93.3 | **94.1** |
| Qwen3.6-27B | 26.90B | Q6_K | 87.9 | 96.9 | 95.4 | 97.3 | 92.7 | **94.0** |
| Qwopus3.6-27B-Coder-Compat | 27.32B | Q6_K | 87.9 | 96.7 | 95.3 | 97.8 | 90.9 | **93.7** |
| Qwable-5-27B-Coder | 26.90B | Q6_K | 87.9 | 97.1 | 95.5 | 97.0 | 90.9 | **93.7** |
| Qwen3.6-35B-A3B | 34.66B | UD-Q4_K_M | 85.0 | 95.7 | 93.3 | 96.7 | 95.7 | **93.3** |
| Qwen3.6-27B | 26.90B | NVFP4 | 87.0 | 96.7 | 94.9 | 97.1 | 90.2 | **93.2** |
| Qwen3-Coder-Next | 79.67B | UD-Q2_K_XL | 83.7 | 96.0 | 89.3 | 96.0 | 93.3 | **91.7** |
| Gemma 4 12B-it | 11.91B | Q6_K | 78.9 | 94.0 | 81.6 | 96.4 | 87.2 | **87.6** |
| gpt-oss-20b | 20.91B | Q4_K_M | 78.6 | 94.6 | 74.5 | 94.8 | 94.5 | **87.4** |
| Nemotron-3-Nano | 31.58B | UD-Q4_K_XL | 74.5 | 89.9 | 75.6 | 90.5 | 80.5 | **82.2** |
| Nemotron-Cascade-2 | 31.58B | Q4_K_M | 74.4 | 91.5 | 75.7 | 87.1 | 79.3 | **81.6** |
| North-Mini-Code-1.0† | 30.48B | Q6_K | 73.3 | 60.2 | 70.8 | 95.8 | 86.6 | **77.4** |

† North-Mini-Code-1.0 is a *reasoning* model, run think-OFF for board parity. ARC-Challenge is reasoning-gated: 60.2 think-OFF to ~95 think-ON (+35), which deflates its q_avg. See the [report](reports/north-mini-code.md).

### Thinking ON (reasoning · extended chain-of-thought)

| Model | Params | Quant | MMLU | ARC-C | HellaSwag | GSM8K | HumanEval | q_avg |
|-------|-------:|-------|-----:|------:|----------:|------:|----------:|------:|
| Qwen3.6-35B-A3B | 34.66B | UD-Q6_K | 94.7 | 97.0 | 87.0 | 92.0 | 98.0 | **93.8** |
| gpt-oss-120B¹ | 116.83B | MXFP4 | 89.5 | 95.0 | 80.0 | 97.0 | 98.0 | **91.9** |
| Qwen3.6-28B-REAP-A3B | 28.24B | Q6_K | 87.7 | 95.0 | 82.0 | 90.0 | 94.0 | **89.7** |

> **HumanEval correction (2026-06-04).** An earlier harness passed API stop sequences (`\ndef`, `\nclass`) that fired *mid-reasoning*, truncating inline-reasoning models before they emitted code — producing false-low scores (Qwen3-Coder-Next read **10%**, not 93%). Every model has since been re-run on the fixed, reasoning-aware harness (no stop sequences, `max_tokens=4096`, indentation-preserving response handling). A second extraction fix (2026-06-04) makes program assembly format-agnostic — it generates candidate assemblies and keeps whichever one compiles — after Nemotron-3-Nano exposed a case where the model indents only the *first* body line differently (raw HumanEval read **21%**; corrected to 80.5%). **Do not cite any HumanEval figure published before this date.**
>
> **Why two tables.** Thinking-off rows answer directly; thinking-on rows emit an extended reasoning chain first. The two modes are not comparable on the same axis — including on MCQ/GSM8K — so they are ranked separately. Within a family, turning thinking on trades raw knowledge recall for reasoning depth (compare Qwen3.6-35B-A3B in both tables: MMLU 85.0 → 94.7).
>
> ¹ **gpt-oss-120B** runs via MoE CPU-offload (`--n-cpu-moe 20`) — it does not fit 32GB VRAM (59GB model); ~30GB VRAM + the rest in system RAM, ~47 tok/s generation. It and the other two thinking-on rows were run on a ~100-item-per-task subset (MMLU 2/subject).

> **Sampling.** MMLU & HellaSwag use 50% stratified sampling (seed=42); ARC-Challenge, GSM8K, and HumanEval run the full item counts (HumanEval = all 164). Full per-model reports in [`reports/`](reports/).

### Methodology

| Benchmark | Dataset | Few-shot | Scoring | Items |
|-----------|---------|----------|---------|------:|
| MMLU | `cais/mmlu` | 5-shot | Letter extraction (A/B/C/D) | 14,042 |
| ARC-Challenge | `allenai/ai2_arc` | 25-shot | Letter extraction | 1,172 |
| HellaSwag | `Rowan/hellaswag` | 10-shot | Letter extraction | 10,042 |
| GSM8K | `openai/gsm8k` | 5-shot CoT | Exact numeric match | 1,319 |
| HumanEval | `openai/openai_humaneval` | 0-shot | pass@1 (code execution) | 164 |

All benchmarks run at `temperature=0`. MCQ and GSM8K use `max_tokens=2048`; HumanEval uses `max_tokens=4096` with **no stop sequences** (reasoning models emit code only after long inline reasoning — premature stops were the bug corrected above). Multiple-choice tasks use generative letter extraction instead of loglikelihood scoring — scores are internally consistent for model comparison but may differ from logprob-based evaluations by 5-15%.

Full per-model reports with MMLU category breakdowns, parse reliability stats, and speed data: [`reports/`](reports/)

---

## Speed Benchmarks

### What's measured

- **Prompt processing (pp)**: parallel batched token throughput at context lengths 128, 512, 2048, 4096, 8192, 16384
- **Text generation (tg)**: sequential autoregressive token throughput at 128 tokens
- All models fully GPU-offloaded (ngl=99)

### Speed data schema

| Column | Description |
|--------|-------------|
| `model` | Model name |
| `architecture` | Dense or MoE (with active param count) |
| `params_b` | Total parameters in billions |
| `quant` | Quantization method |
| `size_gib` | File size in GiB |
| `engine` | Inference engine (llama.cpp or vLLM) |
| `backend` | Compute backend (CUDA) |
| `gpu` | GPU model |
| `vram_gb` | VRAM in GB |
| `test` | Benchmark test (pp128, pp512, ..., tg128) |
| `tokens_per_sec` | Throughput in tokens/second |
| `stddev` | Standard deviation |
| `date` | Benchmark date |

### Key findings

MoE (3B active) vs Dense (27B) on same-family Qwen3.6 models:
- Prompt processing: **2.4x faster** across all context lengths
- Text generation: **3.5x faster** (271 vs 77 t/s)
- Both degrade ~17% at 16K context (attention + VRAM, not parameter count)

---

## Field Reports

One-shot investigations that don't fit the leaderboard format — claim verification, new-architecture probes, and consumer-hardware autopsies, all measured on the same rig. Newest first.

| Report | Finding |
|---|---|
| [FP4 on a consumer 5090: the Blackwell headline loses to plain int4](reports/fp4-consumer-blackwell.md) · [chart](reports/fp4-consumer-blackwell.png) | FP4 is the Blackwell selling point — benchmarked on one RTX 5090 (sm_120) against the quants you'd actually run, Qwen3-14B in vLLM 0.21. Two findings that compound. (1) NVFP4 is the only quant that won't run out of the box: AWQ/FP8 use prebuilt Marlin kernels, but NVFP4 makes FlashInfer JIT-compile native sm_120 FP4 cutlass kernels at load — needing ninja on PATH + a real CUDA toolkit + the correct CUDA_HOME (the default /usr/local/cuda-13.0 doesn't exist on the box) + a flashinfer-cache clear. (2) Once native FP4 is genuinely running (declared modelopt_fp4, not a Marlin dequant fallback), it's still slower than AWQ int4 at every batch: batch-1 decode AWQ 150 vs NVFP4 100 (0.66x) vs FP8 90; batch-32 AWQ 3937 vs NVFP4 3321. The 4-6x FP4 numbers are B200 tensor-core throughput; on consumer sm_120 a mature int4-Marlin kernel wins. bf16-14B doesn't fit 32GB (no KV room). Verdict: use AWQ int4, skip NVFP4 on consumer Blackwell. (QuTLASS / MR-GPTQ's real-FP4 4x claim, arXiv 2509.23202, is the parked follow-up.) |
| [One-Shot EM doesn't reproduce on a 5090: entropy fell, accuracy didn't](reports/one-shot-em.md) · [chart](reports/one-shot-em.png) | One-Shot Entropy Minimization (arXiv 2505.20282) claims +24.7 on Qwen2.5-Math-7B from ONE unlabeled example in ~10 steps, no rewards. The full-param recipe OOMs on 32GB (14GB weights + 14GB bf16 grads), so the consumer-feasible version is LoRA (batch 16). Measured greedy pass@1 with the authors' grader: base reproduces the paper (MATH500 53.4 vs 53.0), but EM adds +2.0 at its peak step then collapses back by step 15, and AMC23 goes −2.5 (claim was +25.8 / +26.2). The keeper is the mechanism: the entropy objective trained fine (mean per-token entropy 0.098→0.035) while accuracy stayed flat — distribution-sharpening, not learning, in the paper's own words. Same base as Spurious Rewards (+21 on MATH500 from random rewards). Honest limits: full-param didn't fit, so this isn't a refutation of the multi-GPU number; the few-shot format control backfired on Qwen-Math's native zero-shot CoT. |
| [Draft-free spec-decode on Qwen3-8B: workload split + dumb-vs-fancy null](reports/cacheback-spec-decode.md) · [chart](reports/cacheback-spec-decode.png) | Cache-only speculative decoding (no draft model, no extra VRAM) on Qwen3-8B: one greedy loop, three drafters (AR / one-line prompt-lookup / Cacheback's dynamic LRU table) × three workloads. The speedup is workload-shaped, not method-shaped — code 1.45-1.47x, summarize 1.30x, open chat 1.26x, with MAT tracking speedup 1:1, biggest where local agents live and no workload at zero. The null: Cacheback's LRU table ties one-line prompt-lookup (identical on 2/3 workloads, +0.01 MAT on code) — at leader-length 1 they are the same algorithm; Cacheback's real edge is its frozen corpus + tree drafting, not the dynamic table, and the cited "1.86x" is Vicuna-7B + frozen corpus on a 4090, not a modern 8B. Lossless: 46072/46080 tokens byte-identical to greedy; the 8 misses are exact bf16 logit ties (gap 0.000) where greedy itself is non-deterministic, not a decoder bug. |
| [Sovereign TTS head-to-head: 1.7B Apache beats 4B research-license](reports/tts-head-to-head.md) · [chart](reports/tts-head-to-head.png) | Qwen3-TTS-1.7B (Apache) vs Fish-S2-Pro (4B, research license) on one RTX 5090, 150 Seed-TTS-eval EN utterances, both bf16 and neither compiled. Round-trip WER is a tie (0.6% each — Fish's sub-1% claim holds, Qwen matches it); SIM-o 0.699 vs 0.625; but RTFx 2.22× vs 0.39× and first-audio latency 1.72s vs 9.74s. Fish-S2-Pro's serving stack assumes SGLang + torch.compile + datacenter cards (its RTF<0.5 is an H200 number) — out-of-the-box on a consumer GPU the small open model is 5.7× faster at the same intelligibility. Size + serving assumptions, not quality. A compiled-Fish rerun is the obvious follow-up. |
| [GLM-5.2 autopsy: clever MLA+DSA, still datacenter-only](reports/glm-5-2-autopsy.md) · [chart](reports/glm-5-2-walls.png) | The biggest open-weights drop in months (743B MoE, MIT, 1M context), measured by arithmetic from the published config — not served. Credit first: `glm_moe_dsa` = MLA + DeepSeek sparse attention compresses the KV cache ~57× (1M = ~88 GiB vs ~4.9 TiB). But the weights can't fit 96GB addressable at any quant (743B needs 1.03 bits/weight; smallest ~1.58-bit = ~147 GB), the 1M KV alone ≈ the whole machine, and DSA saves compute not memory. ~235 GB to use 1M context — past a single H200. Clever ≠ consumer; the home-lab move is to wait for a GLM-5.2-Air. |
| [Qwen3.6-27B pi-tune: the coding tune that works](reports/qwen3-6-27b-pi-tune.md) · [chart](reports/pi-tune-provenance.png) | A community QLoRA SFT of Qwen3.6-27B on REAL non-thinking agent traces, measured controlled vs its base at matched Q6_K across four legs. Quality (93.3 vs 94.0) and synthetic Agentic Score (98.01 vs 98.61) stay flat — but real SWE-bench Verified resolve goes UP 19 → 20/30 (give-ups 8 → 6) and the MTP drafter holds (2.0-2.4× vs base 1.8-2.2×) where Qwopus-Coder's degraded. The first of three Qwen3.6-27B coding tunes to improve real bug-fixing: across all three the synthetic score is a 2.4pt band while real SWE spans 11-20, so training-data provenance (real traces > synthetic distill), not the "agentic coder" label, is what the anchor sees. |
| [Qwable-5-27B-Coder: real traces, still regresses](reports/qwable-5-27b-coder.md) · [chart](reports/qwable-5-anchor.png) | A real-trace SFT of Qwen3.6-27B (Claude Fable-5 then Kimi 2.7 Coder agent traces), measured controlled vs its base at matched Q6_K. Quality (93.7 vs 94.0) AND the synthetic Agentic Score (98.61 vs 98.61, identical to the decimal) stay flat — but real SWE-bench Verified resolve drops 19 → 17/30 and give-ups rise 8 → 10. The 4th Qwen3.6-27B coding tune on the anchor: real traces are necessary but not sufficient (pi-tune's terminal/repo/DevOps traces remain the only data that moved real resolve up). The MTP drafter survived the SFT (1.9-2.4×, base range) — so drafter-survival is not capability-preservation. |
| [Qwable-3.6-27b: the distill every cheap eval passes, SWE-bench fails](reports/qwable-3.6-27b-q4.md) · [chart](reports/qwable-27b-flat-cliff.png) | A dense Qwen3.6-27B + Fable-5-style SFT, measured controlled vs its base at matched Q4_K_M. Quality (93.4 vs 94.0) AND the synthetic Agentic Score (97.64 vs 98.19) stay flat — but real SWE-bench Verified resolve drops 18 → 11/30 and give-ups rise 7 → 13. Quant ruled out (base Q6 → Q4 = −1 bug). The agentic board would call it neutral; only the reality anchor caught the give-up regression — a 3rd failure mode, and the inverse of the MoE Qwable-v1 (whose synthetic honestly declined). |
| [Qwable-v1: agentic distillation regresses vs base](reports/qwable-v1.md) · [chart](reports/qwable-decline.png) | A Claude-Code/Fable-5-distilled "agentic coder" (Qwen3.6-35B-A3B) measured controlled vs its vanilla base + the Opus-reasoning-distill, same Q5_K_M. Every post-train step LOWERS the agentic score (99.58 → 97.92 → 96.25), and real SWE-bench Verified resolve drops 19 → 11/30 with give-ups nearly doubling (9 → 16 empty patches). Not a mirage — synthetic fairly predicts real here; the distillation regressed a top-tier base (the vanilla base ties best real resolve on the board). |
| [ECHO maze microcosm](reports/echo-maze-microcosm.md) · [chart](reports/echo-maze-microcosm.png) | ECHO's "world model for free" env-token loss, isolated in a 10M from-scratch maze transformer (behavior cloning; the only A/B is the loss mask). A clean null under both observation regimes — full-obs and walls-only — every gap inside the seed spread, no growth with maze size. The free aux loss is worth exactly that in pure imitation; the reported gain must live in ECHO's on-policy RL coupling, not the loss as a plug-in. (The paper has no maze — this microcosm is original, verified by grepping the source + repo.) |
| [Spec-decode three-way (Gemma 4 26B-A4B)](reports/specdecode-gemma-4-26b-a4b.md) · [chart](reports/specdecode-gemma-4-26b-a4b.png) | MTP vs EAGLE-3 vs DFlash on one RTX 5090 (vLLM 0.21, sm_120), including the EAGLE-3 leg nobody publishes. Single-stream near-tie: DFlash 2.19x, MTP 2.13x, EAGLE-3 1.69x — DFlash is feast-or-famine (prose 1.04x, repetitive 4.37x), MTP the steady all-rounder. Six consumer-Blackwell fixes to run it at all (NVFP4 to MARLIN, FLEX_ATTENTION for the #42068 attention deadlock) plus a self-caught /metrics parser bug. Dense 31B excluded: no clean quant fits 32GB. |
| [Qwopus3.6-27B-Coder](reports/qwopus-coder-27b.md) · [chart](reports/qwopus-coder-chart.png) | Four legs measured: q_avg 94.1 (#2 thinking-off, beats its base at a smaller quant); "100 tps" MTP verified (96-114 t/s) but the finetuned head accepts worse than the original (1.4-1.6x vs 1.8-2.2x); a perfect 100 Agentic Score — and 57% real SWE-bench resolve, *below its own base* (63%). Trained on Hermes traces: the in-distribution mirage the reality anchor was built to catch. 67% claim doesn't reproduce. |
| [Qwopus3.6-27B-Coder-Compat: the regressions healed](reports/qwopus-coder-compat.md) · [chart](reports/qwopus-coder-compat-recovery.png) | The "compatibility" re-release of Qwopus-Coder, measured controlled vs base Qwen3.6-27B + the prior Coder at matched Q6_K (think-off, temp 0). Quality flat (q_avg 93.7). The prior tune's two regressions both heal: the degraded MTP draft head recovers 1.4-1.6x → 1.9-2.3x (back on the base curve), and real SWE-bench Verified resolve returns to base parity — 8/12, the exact same bugs as base, +1 over the prior tune (recovers pytest-6202) with one fewer give-up. Agentic Score 100.0 ties the prior (saturated — the synthetic axis can't separate the two; the anchor can). A compat fix that costs no capability: the coder tune no longer carries a drafter or real-bug penalty. |
| [Keye-VL-2.0-30B autopsy](reports/keye-vl-2-autopsy.md) · [chart](reports/keye-walls-chart.png) | Five measured walls: "lossless 256K" needs 25.8GB of KV alone; the shipped sparse attention is O(N²)-memory (one 30.65GiB allocation at ~32K, measured); 4-bit quant reaches 4.7% of params; the code's API window is two transformers release candidates wide. Does not run on consumer hardware. |
| [LocateAnything-3B](reports/locateanything-3b.md) · [chart](reports/la-screenspot-chart.png) · [raw](raw/la-screenspot.jsonl) | ScreenSpot-Pro 55.3% measured vs 60.3 claimed (32GB forces extra downscale; accuracy tracks screenshot size). Real fault line: text 63.2% vs icons 42.7%. PBD parallel box decoding verified at 2.07x on the SDPA fallback. |
| [HRM-Text-1B](reports/hrm-text-1b.md) · [chart](reports/hrm-trap.png) · [raw gens](raw/) | GSM8K 79.5% (claimed 84.5: holds at n=200). Omitting `token_type_ids` — which every standard harness does — silently costs 26 points. The recurrence bill: a 1.2B that decodes like a ~5B (42.9 tok/s bf16, 4x KV cache). |
| [DiffusionGemma vs AR](reports/diffusion-vs-ar.md) · [chart](reports/diffusion-vs-ar.png) | AR wins at every answer length: diffusion pays a fixed ~3s per 256-token canvas (0.8 effective tok/s on short answers; best case still 2.3x slower). Day-0 public GGUFs were unloadable — convert from source. |
| [Gemma 4 31B QAT + MTP](reports/gemma4-31b-qat-mtp.md) · [chart](reports/gemma4-qat-speed.png) | The MTP draft head lifts decode 76 to 125 tok/s (1.67x). QAT's real value is VRAM, not quality: the Q4 footprint is what fits 128K context plus the draft head on one card. |
| [NVFP4 vs Q6_K](reports/nvfp4-vs-q6-qwen3-6-27b.md) · [chart](reports/chart-nvfp4-vs-q6.png) | Qwen3.6-27B: NVFP4 trades ~1pt q_avg against Q6_K. |
| [GRPO on one 5090](reports/fp8-rl-grpo.md) · [chart](reports/chart-grpo-gsm8k.png) | Single-GPU RL: +7.66 GSM8K on a 4B. Train-prompt-to-eval-prompt alignment is the lever. |
| [Embedding retrieval bench](reports/embedding-retrieval-bench.md) · [chart](reports/embed-bench-scatter.png) | Local embedding models benchmarked for retrieval quality vs speed on the 5090. |
| [Mistral Small 4 speed](reports/mistral-small-4-speed.md) · [chart](reports/mistral-vs-gptoss-speed.png) | Speed profile vs gpt-oss-20b — and a benchmarking trap: reasoning is gated behind `reasoning_effort`, which defaults off. |
| [LFM2.5-VL 1.6B extraction](reports/lfm2-5-vl-1-6b-extract.md) · [chart](reports/chart-lfm2-vl-extract.png) | A 1.6B VL model as a local structured-data extractor. |
| [Nex-N2-mini agentic probe](reports/agentic-nex-n2-mini.md) | Adaptive Thinking saves 65% of tokens but costs 13pts task success. Superseded by the dedicated Agentic Score leaderboard (below). |

## Related Datasets

- [witcheer/agentic-score-leaderboard](https://huggingface.co/datasets/witcheer/agentic-score-leaderboard) — model-agnostic agentic tool-calling benchmark (7 models, 40 tasks) + the SWE-bench reality anchor
- [witcheer/sovereign-asr-bench](https://huggingface.co/datasets/witcheer/sovereign-asr-bench) — local ASR on the 5090: Parakeet-TDT vs Whisper (WER / RTFx / VRAM)

---

## Hardware

| Component | Spec |
|-----------|------|
| GPU | NVIDIA GeForce RTX 5090 32GB (Blackwell, sm_120a) |
| CPU | AMD Ryzen 5 9600 (6c/12t) |
| RAM | 64GB DDR5-5600 |
| OS | Ubuntu 26.04 LTS |
| CUDA | 12.8 (patched for glibc 2.41) |

## Tooling

All benchmarks generated with [llm-bench-rig](https://github.com/notwitcheer/llm-bench-rig) — open-source pipeline for speed and quality benchmarks on GGUF and safetensors models.