HAIL.SO / DISPATCH · 2026-06-20 · COMPARE
FILE: COMPARE · 6 of 6 slots

side by side.

← all costsclear
01

Large language models

2 models
Mistral
Codestral 25.08
×
Perplexity
Sonar
×
Output $/MTok$0.90$1.00
Input $/MTok$0.30$1.00
Cached input $/MTok
Context window262,144128,000
Output cap8,1928,000
Modalitiesin: text / out: textin: text / out: text
Tool use
Structured output
FamilyCodestralSonar
Knowledge cutoff
Verified2026-05-18stale 33d2026-05-19stale 32d
Sourceprovider page ↗provider page ↗
02

Speech to text

2 models
AssemblyAI
Universal-2
×
AssemblyAI
Universal-Streaming
×
$/min$0.002500$0.002500
$/min batch
Streaming
Realtime
Languages99+en
Diarizationextra-costextra-cost
TTFW
Verified2026-05-05stale 46d2026-05-05stale 46d
Sourceprovider page ↗provider page ↗
03

Text to speech

2 models
Deepgram
Deepgram Aura 2
×
Smallest.ai
Lightning v3.1
×
$/1M chars$30.00$25.00
Voice qualityneuralneural
Voice cloning✓ included
Voice count217
Languagesen, es, de, fr, nl, it, jaen, hi, es, mr, kn, ta, bn, gu, te, ml, pa, or
SSML support
TTFB200ms
Output formatswav, mp3, linear16, mulaw, alaw, opuspcm, wav, mp3, mulaw
Verified2026-05-19stale 32d2026-05-19stale 32d
Sourceprovider page ↗provider page ↗