Home / LLM launch watch
Starting snapshot of publisher-sourced model launch records. Announcement date is the publisher-stated date. First seen is when this record was checked — not when the model launched. Historical announcement pricing is labeled historical and is not a complete comparison.
No schedule, cron, webhook or standing automation is attached to this data. This is a starting snapshot, not complete launch coverage. No match does not mean no launch occurred.
Mistral's changelog lists zai-glm-5-3-flash: multimodal 320B total / 18B active parameters, 1M-token context. License not reviewed in this launch record.
Changelog price at announcement, not reverified checkout pricing. Not a complete provider comparison.
Publisher announces availability on Claude Platform, AWS, Google Cloud and Azure. Prices below are Anthropic's launch table, not verified prices for those other platforms. Parameter count not disclosed in the inspected announcement. Hosted model; no open-weight license inferred.
Publisher launch table, historical announcement pricing. Not a complete provider comparison.
Publisher launch table, historical announcement pricing. Not a complete provider comparison.
1T total / 52B active parameter multimodal model. Changelog lists 1M context. The Oct 6 announcement says weights will arrive by month end; this record does not establish that they are released now or license terms. Two-week launch promotion is not a current price.
No price shown: a verified current per-token figure was not established for this record.
API model id: gpt-6.1-terra. Prices below are from the OpenAI pricing page, re-read 2026-10-28. Cached input is a separate SKU at $0.05 / 1M. Hosted model; no open-weight license inferred.
Publisher pricing page, re-read 2026-10-28. Not a complete provider comparison.
Gemini API model. Publisher states an introductory price valid through 2026-11-30, after which standard rates apply. Cached input is stated as 95% off the input rate; the exact per-token figure is not printed on the announcement page and is not shown here. Hosted model; no open-weight license inferred.
Cache read: 95% off input rate — exact figure not printed on announcement page
Publisher announcement page, re-read 2026-10-28. Introductory price; expires 2026-11-30. Not a complete provider comparison.
Publisher announcement page, re-read 2026-10-28. Post-introductory rate stated on announcement page. Not a complete provider comparison.
Prices below are from the Anthropic pricing page, re-read 2026-10-28. Hosted model; no open-weight license inferred.
Publisher pricing page, re-read 2026-10-28. Not a complete provider comparison.
Prices below are from the Anthropic pricing page, re-read 2026-10-28. Hosted model; no open-weight license inferred.
Publisher pricing page, re-read 2026-10-28. Not a complete provider comparison.
740M-parameter multimodal embedding model, 8K context. The publisher announcement states Apache 2.0 license; the license file itself was not read and this record does not constitute a legal determination. No price is published on the announcement page.
No price shown: a verified current per-token figure was not established for this record.
Five audio models: voxtral-mini-tts-3-0, voxtral-mini-tts-realtime-3-0, voxtral-mini-transcribe-3-0, voxtral-mini-transcribe-realtime-3-0, voxtral-small-transcribe-3-0. These are audio models with native per-minute and per-character pricing; prices are not per token and are not converted. Prices re-read from Mistral pricing page 2026-10-28.
Publisher pricing page, re-read 2026-10-28. Native per-minute unit; not per token.. Not a complete provider comparison.
Publisher pricing page, re-read 2026-10-28. Native per-minute unit; not per token.. Not a complete provider comparison.
Publisher pricing page, re-read 2026-10-28. Native per-minute unit; not per token.. Not a complete provider comparison.
Publisher pricing page, re-read 2026-10-28. Native per-character unit; not per token.. Not a complete provider comparison.
Publisher pricing page, re-read 2026-10-28. Price not listed on Mistral pricing page.. Not a complete provider comparison.
Two small API models released together. Prices below are as printed on the publisher announcement page, re-read 2026-10-30. GPT-5.4 mini has a 400k context window and is available in the API, Codex, and ChatGPT. GPT-5.4 nano is available in the API only. API model ids are not printed on the announcement page. Hosted models; no open-weight license inferred.
Publisher announcement page, re-read 2026-10-30. Not a complete provider comparison.
Publisher announcement page, re-read 2026-10-30. Not a complete provider comparison.