Side by side

See the difference.

Scores, capabilities and costs for the models you’re considering.

Scroll to compare models →

Comparison of nvidia/canary-1b-v2, Elorian: Glimpse
2 modelsnvidianvidia/canary-1b-v2Remove ×elorianElorian: GlimpseRemove ×
What’s different1B parameters · 25 languages · Encoder: Conformer-based · Decoder: TransformerGlimpse is a multimodal visual reasoning model from Elorian, built for visual understanding and complex visual reasoning in the physical world. It reasons directly over images and video rather than...
DateOct 17, 2026DiscoveredOct 17, 2026Listed
Listed rateSee provider quotes$0.60 in / $3.00 out/ 1M tokensTogether · via OpenRouter · all routes ↗together
Provider endpoints0 endpoints →2 endpoints →
AMI-Cleaned WER ↗English short-form; normalization unversioned13.03% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
Earnings22-Cleaned-AA-chunked WER ↗English short-form; normalization unversioned7.28% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
Gigaspeech-Cleaned WER ↗English short-form; normalization unversioned9.12% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
LS Clean WER ↗English short-form; normalization unversioned1.76% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
LS Other WER ↗English short-form; normalization unversioned3.07% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
SPGISpeech WER ↗English short-form; normalization unversioned2.03% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
URGENT2024 WER ↗English short-form; normalization unversioned14.56% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
Voice Arena Monsoon WER ↗English short-form; normalization unversioned4.97% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
Voxpopuli-AA-Cleaned WER ↗English short-form; normalization unversioned4.39% WER ↗nvidia/canary-1b-v2Leaderboard entryNot reported
Context window—262.1K
Max output—256K
Parameters1B—
Tool callingNot reportedSupported
Structured outputNot reportedSupported
Image inputNot reportedSupported
Audio inputSupportedNot reported
TranscriptionSupportedNot reported
Video inputNot reportedSupported
ReasoningNot reportedSupported
Prompt cachingNot reportedSupported

Add another model

Search all models →