facffbc6c8
AIX: require clean connection ends. #980
2026-02-12 16:31:41 -08:00
dd5b7cb8c2
AIX: dispatch: increase debugging vendor-initiated disconnect. #980
2026-02-12 14:12:25 -08:00
3dc61109d7
AIX: Server: debug recovered packets
2026-02-12 01:34:10 -08:00
9ef84260b0
Z.ai: no bits
2026-02-11 22:09:54 -08:00
a4adce5c79
Z.ai: AIX: fix reasoning effort
2026-02-11 17:43:23 -08:00
9e4174df53
Z.ai: AIX: fix dispatch
2026-02-11 17:36:04 -08:00
b5975713a3
Z.ai: OCR does not support WebP
2026-02-11 17:26:23 -08:00
0cd04266b7
Z.ai: improve model spec
2026-02-11 17:26:23 -08:00
5cbd162454
Z.ai: Reasoning settings support
2026-02-11 17:26:23 -08:00
bea1600358
AIX: OpenAI ChatCompletions: empty reasoning_content yields to non-empty content
2026-02-11 17:26:22 -08:00
6a2e201cf5
Z.ai: discovered + curated models support
2026-02-11 17:26:22 -08:00
960551933e
Z.ai LLM vendor support
...
Note: we don't include server-side config anymore starting from this. To stress test the config system.
2026-02-11 17:26:22 -08:00
8b38b6416d
Z.ai: icon & sprite
2026-02-11 17:26:22 -08:00
fac4c39f48
Fix copying of message Sources. Fixes #977 . Fixes #978 .
2026-02-11 13:02:32 -08:00
4912a03250
LLMs: Anthropic: Fast mode research preview
2026-02-10 13:22:47 -08:00
3b13580613
LLMs: parameter-value-based enum price multipliers
2026-02-10 13:04:05 -08:00
95905113ac
LLMs: cached isLLMChatFree_cached
2026-02-10 12:17:21 -08:00
e5387c2323
AIX: Moonshot: remove empty messages
2026-02-10 11:07:09 -08:00
49b61495d0
LLMs: Vendor Settings: unbreak hide advanced despite initially in CSF. Fixes #969
2026-02-09 23:30:41 -08:00
83161bbe98
AIX: Anthropic: Parser: hotfix for 4.6 to elide the double-newline at the beginning when present
2026-02-09 04:50:14 -08:00
4b166120e6
AIX: Anthropic: Dispatch: hotFix for 4.6 interleaved reasoning blocks back-to-back
2026-02-09 04:23:28 -08:00
04494ac752
AIX: Anthropic: Dispatch: hotFix for empty text blocks produced by 4.6 - incoming from the Anthropic API
2026-02-09 04:23:27 -08:00
979809ddb1
AIX: Anthropic: Parser: rename hotFix
2026-02-09 04:23:26 -08:00
5d797c3339
AIX: Anthropic: warn if blocks come out of order, now that Anthropic has fixed it
2026-02-09 04:22:35 -08:00
c337b70a42
LLMs: Anthropic: copy
2026-02-09 01:40:11 -08:00
4bf621f128
LLMs: OpenAI GPT-5.3-Codex speculative support
2026-02-07 13:42:12 -08:00
33505dbb8e
LLMs: Anthropic/OpenRouter: align behavior, align UI #962
2026-02-06 22:40:55 -08:00
c81e1f144f
AIX: OpenRouter: protocol bits
2026-02-06 20:56:39 -08:00
38ac8733f6
AIX: OpenRouter: comment on debug: too risky
2026-02-06 20:10:48 -08:00
737a20ee06
AIX: OpenRouter: enable the stricter 'require_parametrs' mode. #948
2026-02-06 20:05:05 -08:00
19f48b8001
AIX: OpenRouter: wires for OR debug parameters
2026-02-06 19:51:50 -08:00
2dc7ba72b3
AIX/LLMs: bits
2026-02-06 19:30:18 -08:00
e12279dab0
AIX: Anthropic: show the US inference setting when on
2026-02-06 19:24:04 -08:00
2e0c79cb64
LLMs: OpenRouter: also inherit the initial temperature from upstreams
2026-02-06 19:19:33 -08:00
aa697edb8c
AIX: Anthropic: minor API changes
2026-02-06 19:18:54 -08:00
c72e3c58dd
AIX: Anthropic: allow US servers
2026-02-06 19:17:01 -08:00
1de30c8bd5
AIX: Anthropic: accomodate some API changes
2026-02-06 18:52:58 -08:00
b7fd0bdba7
LLMs: OpenRouter: auto-inherit configurable parameters from Anthropic, Gemini and OpenAI.
...
Fixes #948 : OpenAI-through-OR verbosity is sync'd with OpenAI models.
Fixes #893 : Gemini-through-OR parameters are synchronized with Gemini models
Fixes #940 : OpenAI-through-OR reasoning effort is synced with OpenAI models and much improved. We will have to still fix #944 for OpenAI levels to be fully sync'd with upstream (in progress)
2026-02-06 18:27:38 -08:00
58457cac50
LLMs: OR/Anthropic: support effort and adaptive.
...
Fixes #962
2026-02-06 18:27:38 -08:00
0fbacee7dc
LLMs: Anthropic: editable Max effort. #962
2026-02-06 18:27:38 -08:00
a498f28d14
LLMs: Anthropic: support for max effort. #962
2026-02-06 18:26:07 -08:00
5b9c6a2d0e
LLMs: Anthropic: support adaptive thinking correctly. #962
2026-02-06 18:26:07 -08:00
4c7f50ab98
LLMs: Anthropic: inline thinking budget
2026-02-06 18:26:07 -08:00
ef03d33bbf
LLMs: Anthropic: GA skills
2026-02-06 18:26:07 -08:00
c952fd734f
LLMs: Opus 4.6: remove forcing
2026-02-06 18:26:07 -08:00
310e99af23
LLMs: Opus 4.6: sort order, unhide 4.5
2026-02-06 18:26:07 -08:00
61a60c5b9f
Markdown: bundle in main chunk instead of lazy-loading
2026-02-06 12:41:41 -08:00
6f4fabf147
Claude Opus 4.6 baseline support
2026-02-05 12:02:21 -08:00
748991249a
LLMs: OpenAI: Update tooling availabiltiy across models
2026-02-05 02:36:28 -08:00
e586142190
AIX: OpenAI-compatible: ChatCompletions: support verbosity for all (not just openrouter)
2026-02-05 00:07:36 -08:00
Enrico Ros