Side by side
See the difference.
Scores, capabilities and costs for the models you’re considering.
Scroll to compare models →
| 3 models | nvidianvidia/canary-1b-v2Remove × | reflectionReflection: Beam 1.0 501B A23BRemove × | moondreamMoondream: Moondream 3 PreviewRemove × |
|---|---|---|---|
| What’s different | 1B parameters · 25 languages · Encoder: Conformer-based · Decoder: Transformer | Apache-2.0 text model with 501B total and 23B active parameters; ships BF16, FP8 and NVFP4 weights. | Moondream 3 Preview is a vision-language model from Moondream with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over… |
| Date | Oct 17, 2026Discovered | Oct 16, 2026Released | Oct 15, 2026Listed |
API prices USD · provider configuration shown below | |||
| Listed rate | See provider quotes | $0.25 in / $1.00 out/ 1M tokensNebius · via OpenRouter · all routes ↗nebius/fp8 · quantization=fp8 | $0.30 in / $2.50 out/ 1M tokensPrimeIntellect · via OpenRouter · all routes ↗primeintellect/fp8 · quantization=fp8 |
| Provider endpoints | 0 endpoints → | 6 endpoints → | 2 endpoints → |
Benchmarks Each result retains its evaluation configuration | |||
| AA Intelligence ↗4.3.2 | Not reported | 30.33 ↗Effort not specifiedIndependent result | Not reported |
| AA-Briefcase ↗1.1 | Not reported | 976.2 ↗Effort not specifiedIndependent result | Not reported |
| GDPval-AA ↗2.1 | Not reported | 1,177.08 ↗Effort not specifiedIndependent result | Not reported |
| AutomationBench-AA ↗AA Index 4.3.2 collection | Not reported | 48.46% ↗Effort not specifiedIndependent result | Not reported |
| Terminal-Bench 4.0 · AA ↗ | Not reported | 3.03% ↗Effort not specifiedHarness unreportedIndependent result | Not reported |
| SciCode ↗AA Index 4.3.2 collection | Not reported | 48.5% ↗AA · Under reviewEffort not specifiedIndependent result | Not reported |
| Humanity’s Last Exam ↗AA Index 4.3.2 collection | Not reported | 34.66% ↗Effort not specifiedIndependent result | Not reported |
| GDP.pdf ↗AA Index 4.3.2 collection | Not reported | 11.6% ↗Effort not specifiedIndependent result | Not reported |
| CritPt ↗AA Index 4.3.2 collection | Not reported | 13.14% ↗AA · Under reviewEffort not specifiedIndependent result | Not reported |
| AA-Omniscience ↗AA Index 4.3.2 collection | Not reported | 1.5 ↗Effort not specifiedIndependent result | Not reported |
| AA-LCR ↗1.1 | Not reported | 78% ↗Effort not specifiedIndependent result | Not reported |
| AMI-Cleaned WER ↗English short-form; normalization unversioned | 13.03% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| Earnings22-Cleaned-AA-chunked WER ↗English short-form; normalization unversioned | 7.28% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| Gigaspeech-Cleaned WER ↗English short-form; normalization unversioned | 9.12% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| LS Clean WER ↗English short-form; normalization unversioned | 1.76% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| LS Other WER ↗English short-form; normalization unversioned | 3.07% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| SPGISpeech WER ↗English short-form; normalization unversioned | 2.03% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| URGENT2024 WER ↗English short-form; normalization unversioned | 14.56% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| Voice Arena Monsoon WER ↗English short-form; normalization unversioned | 4.97% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| Voxpopuli-AA-Cleaned WER ↗English short-form; normalization unversioned | 4.39% WER ↗nvidia/canary-1b-v2Leaderboard entry | Not reported | Not reported |
| LiveBench ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 55.5% ↗Effort: max · single run · equal mean of 7 suitesIndependent result | Not reported |
| LiveBench Design ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 67.9% ↗Effort: max · single runIndependent result | Not reported |
| LiveBench Research ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 72.8% ↗Effort: max · single runIndependent result | Not reported |
| LiveBench Browser Use ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 0% ↗Effort: max · single runIndependent result | Not reported |
| LiveBench Coding ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 83.9% ↗Effort: max · single runIndependent result | Not reported |
| LiveBench Agentic ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 17.3% ↗Effort: max · single runIndependent result | Not reported |
| LiveBench Data Analysis ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 66.4% ↗Effort: max · single runIndependent result | Not reported |
| LiveBench Reasoning ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | 80.2% ↗Effort: max · single runIndependent result | Not reported |
Specifications | |||
| Context window | — | 1M | 32K |
| Max output | — | 943.7K | 4.1K |
| Parameters | 1B | 501B | — |
| Tool calling | Not reported | Supported | Not reported |
| Structured output | Not reported | Supported | Not reported |
| Image input | Not reported | Not reported | Supported |
| Audio input | Supported | Not reported | Not reported |
| Transcription | Supported | Not reported | Not reported |
| Reasoning | Not reported | Supported | Supported |
| Prompt caching | Not reported | Supported | Not reported |
| Open weights | Not reported | Supported | Supported |