GigaAM-v3
GigaAM-v3 is a Conformer-based foundation model with 220–240M parameters, pretrained on diverse Russian speech data using the HuBERT-CTC objective. It is the third generation of the GigaAM family and provides state-of-the-art performance on Russian ASR across a wide range of domains.
GigaAM-v3 includes the following model variants:
ssl— self-supervised HuBERT–CTC encoder pre-trained on 700,000 hours of Russian speechctc— ASR model fine-tuned with a CTC decoderrnnt— ASR model fine-tuned with an RNN-T decodere2e_ctc— end-to-end CTC model with punctuation and text normalizatione2e_rnnt— end-to-end RNN-T model with punctuation and text normalization
GigaAM-v3 training incorporates new internal datasets: callcenter conversations, speech with background music, natural speech, and speech with atypical characteristics.
the models perform on average 30% better on these new domains, while maintaining the same quality as previous GigaAM generations on public benchmarks.
The table below reports the Word Error Rate (%) for GigaAM-v3 and other existing models over diverse domains.
| Set Name | V3_CTC | V3_RNNT | T-One + LM | Whisper |
|---|---|---|---|---|
| Open Datasets | 3.0 | 2.6 | 5.7 | 12.0 |
| Golos Farfield | 4.5 | 3.9 | 12.2 | 16.7 |
| Natural Speech | 7.8 | 6.9 | 14.5 | 13.6 |
| Disordered Speech | 20.6 | 19.2 | 51.0 | 59.3 |
| Callcenter | 10.3 | 9.5 | 13.5 | 23.9 |
| Average | 9.2 | 8.4 | 19.4 | 25.1 |
The end-to-end ASR models (e2e_ctc and e2e_rnnt) produce punctuated, normalized text directly.
In end-to-end ASR comparisons of e2e_ctc and e2e_rnnt against Whisper-large-v3, using Gemini 2.5 Pro as an LLM-as-a-judge, GigaAM-v3 models win by an average margin of 70:30.
For detailed results, see metrics.
FP8 quantization
The e2e_ctc, ctc, e2e_rnnt, and rnnt branches additionally carry FP8 (E4M3) quantized weights alongside the original fp16 weights:
model.safetensors— original fp16 weightsmodel_fp8.safetensors— FP8 E4M3 weights (per-output-channel scales) + per-tensor activation scales (model_fp8.safetensors.activation_scales.json)
Quantization targets the GEMM layers (encoder feed-forward and attention projections; RNNT joint enc/pred). All variants use post-training quantization (PTQ) with per-tensor activation calibration — no fine-tuning is required. FP8 PTQ tracks the fp16 model closely for both CTC and RNNT.
Measured over 1000 held-out audio samples, FP8 transcription closely tracks the fp16 baseline — transcripts are identical for 93–99% of samples, and FP8 WER vs ground truth stays within ±0.2% of fp16:
| Variant | Word disagreement (FP8 vs fp16) | Transcripts identical | ΔWER vs fp16 |
|---|---|---|---|
e2e_ctc | 1.55% | 93.6% | +0.00 |
ctc | 1.59% | 93.4% | +0.06 |
e2e_rnnt | 0.85% | 97.1% | −0.16 |
rnnt | 0.26% | 99.1% | +0.06 |
License: MIT
Paper: GigaAM: Efficient Self-Supervised Learner for Speech Recognition (InterSpeech 2025)