Repository navigation
feat: register Fable 5.2, GPT-6.2 Sol and GPT-6.2 Astra models - #46
Conversation
Adds five registry entries (claude-fable-5-2, global./us.anthropic.claude-fable-5-2, openai.gpt-6.2-sol, openai.gpt-6.2-astra) and three aliases, so Supernova can price these models and read their efforts and limits. The GPT-6.2 entries list text and image input only, since mantle rejects PDF input for both. Co-authored-by: Sam Boland <sam.boland@appfolio.com>
|
Review of 5a39210. The registry change is correct: I re-ran the insertion check against
Verified and fine:
|
I've queued the remaining work (entry 20447) and the coder is on it. It covers these steps:
I haven't run anything yet, so none of this is confirmed. |
Adds both 6.2 ids to the GPT-6 mantle table and checks effort none against 6.1 Sol, 6.2 Sol and 6.2 Astra. Both 6.2 models return no reasoning items in the tool loop, as 6.1 Sol does, so tool_loop is false for them.
Pushed Changed 15 files (+3049 −6): — via Coder session (view) |
|
Review of f2fded9. The code and cassettes are correct. Three statements in the PR description are wrong. Please correct them in your reply so the description regenerates:
Verified and fine:
|
Here are the three corrections, which replace the matching statements in the description:
|
samuelboland
left a comment
There was a problem hiding this comment.
Verified at f2fded9: registry insertion only (1243 -> 1248), 14 cassettes with credentials filtered, mantle spec traits and effort-none table, CI green.
|
🤖 Supernova Code Review — View trace — via automation "Code Review Opus 5.5 Medium" (run), installed by Sam Boland (sam.boland@appfolio.com) |
There was a problem hiding this comment.
Reviewed PR #46 at f2fded9 and found nothing to fix. No inline comments.
What I checked:
models.jsonentries: the three Fable 5.2 entries follow the Fable 5.1 pattern. The anthropic entry lists text, image and pdf input plus citations. The bedrock entries list text and image only, and theus.entry adds streaming. Prices are internally consistent:- The cache read is 0.025x input on all three ids (8 → 0.20, 8.8 → 0.22).
- The cache write is 1.25x input.
- The
us.rates are a 10% premium over the global rates.
- GPT-6.2 prices: both use a cache read of 0.05x and a cache write of 1.25x, the same ratios as 6.1 Sol.
- GPT-6.2 context window: the 1,050,000 window matches the probe in the spec header. Inputs of 904,217 were accepted and about 935,000 rejected, which brackets the expected 922,000 input limit (1,050,000 minus 128,000 reserved for output).
- Aliases:
claude-fable-5-2maps to anthropic and tous.on bedrock, the same as 5.1.gpt-6.2-solandgpt-6.2-astramap to bedrock. responses_gpt6_mantle_spec.rb:- Every example that runs for the 6.2 models has a matching cassette, 12 in total.
- The tool-loop example is skipped for 6.2 (
tool_loop: false), so its missing cassettes don't matter. - The effort-none table keeps the original example name for 6.1 Sol, so its existing cassette still replays. The two new 6.2 cassettes are present.
- Fable 5.2 specs: they look up models by raw id, so the bedrock region prefix isn't changed during lookup. The shared context
with configured RubyLLMresetsbedrock_regionbefore each example, so the spec doesn't leak config into other tests. - Routing spec lists: the 6.2 ids added to
mantle_responses_spec.rbandbedrock_spec.rbmatch the existing GPT-6 pattern.
— via automation "Code Review Opus 5.5 Medium" (run), installed by Sam Boland (sam.boland@appfolio.com)
What this does
Adds three newly available models to the fork's catalog (PR #46, NewModels W1-01).
mainis the 1.16 line that Supernova runs today.No
lib/code changed. Routing already matches the new ids.Providers::Bedrock::MANTLE_ONLY_MODEL_PATTERN,Protocols::MantleResponses::FRONTIER_GPT_PATTERNandProtocols::Responses::Chat::GPT6_MODEL_PATTERNalready matchopenai.gpt-6.2-solandopenai.gpt-6.2-astra. Fable 5.2 uses the same Converse and InvokeModel paths as Fable 5.1. The registry entries, aliases and specs encode live-probe results from 2026-11-18 (account 436292035443).Commit
5a39210: registry and offline specslib/ruby_llm/models.jsongains five entries:claude-fable-5-2(anthropic),global.anthropic.claude-fable-5-2andus.anthropic.claude-fable-5-2(bedrock), andopenai.gpt-6.2-solandopenai.gpt-6.2-astra(bedrock).inserted 5 entries; 1243 -> 1248.models.jsonwas not reformatted or regenerated.aa95f613. Five entries were added and no existing entry changed.lib/ruby_llm/aliases.jsongainsclaude-fable-5-2(anthropic and bedrock),gpt-6.2-solandgpt-6.2-astra.us.id: 8.8 / 44 / 0.22 / 11.offer-agsprgw6z3dg6, read 2026-11-11. The cache read is 0.025x the base input price, as for Fable 5.1.none.offer-7vsvjk6xmsl7c, Astraoffer-i5wczxyjg7weu.textandimageonly. The 6.1 entries also listpdf, but both 6.2 models return 400 on PDF input.spec/ruby_llm/models_fable_5_2_spec.rbchecks prices, efforts and limits for the three ids. It also checks that theclaude-fable-5-2alias resolves to anthropic and that theus.id resolves on bedrock with an input price of 8.8.models_bedrock_regional_pricing_spec.rbadds theus.id and a note that the 0.025x cache read also holds for Fable 5.2.models_gpt_6_bedrock_spec.rbadds both 6.2 ids and aliases, plus an example that their input is text and image only. Its header notes the context window: 904,217 tokens accepted, about 935,000 rejected.mantle_responses_spec.rbandbedrock_spec.rbadd both 6.2 ids to their routing id lists.Commit
f2fded9: mantle spec and cassettesspec/support/rubyllm_configuration.rb:17-20readsAWS_ACCESS_KEY_ID,AWS_SECRET_ACCESS_KEYandAWS_SESSION_TOKENfrom the environment and falls back to'test', and those were unset. The cassettes were then recorded through/tmp/with_creds.rb, which reads the ECS container credentials endpoint and execs the command.spec/ruby_llm/protocols/responses_gpt6_mantle_spec.rbgets two newgpt6_modelsrows for 6.2 Sol and 6.2 Astra. It also gets a table-driven "rejects effort none on " example over 6.1 Sol, 6.2 Sol and 6.2 Astra. The expected errors are:'none' is not supporteddoes not support reasoning effort 'none'Supported values are: 'low', 'medium', 'high', 'xhigh', and 'max'spec/fixtures/vcr_cassettes/: six per 6.2 model, plus one effort-none rejection per model. The seventh per-model example (the tool loop) is disabled, as explained below, so there are not seven per model as in PR GptTerra W1-01: Register GPT-6.1 Terra in the fork's Bedrock catalog #39.Things a reviewer should know
tool_loop: false).replays every reasoning item through a multi-round tool loopis gated byif traits[:tool_loop]and does not run for them.expect_reasoning_replayed(responses_gpt6_mantle_spec.rb:69) withexpected 0.positive? to be truthy, got false: no reply in the loop carried a reasoning item. 6.1 Sol already hadtool_loop: falsefor the same reason.sol: trueis kept for both 6.2 models. That trait only affects the commentary check inside the gated tool-loop example.noneandminimalreturn 400. The lowest valid effort islow.NOT_AVAILABLE, the same as earlier GPT models on mantle.grep -lE 'ASIA[A-Z0-9]{12}|AKIA[A-Z0-9]{12}' spec/fixtures/vcr_cassettes/*6_2*printed nothing. Session tokens are filtered from the cassettes.(gpt6_models.keys - [model]).firstas the "other" model, which isopenai.gpt-6-luna. So the 6.2 cassettes include a call to Luna.thinking.typeofdisabledorenabledreturns 400.temperatureandtop_preturn 400.tool_choiceoftooloranyreturns 400.autoworks.max_tokensof 128,001 is rejected, and so is a prompt over 1M tokens.us.profile is active in us-east-1, us-east-2 and us-west-2.Verification
bedrock_spec.rbgave 43 examples and 0 failures on the same seeds afterrubocop -a. Rubocop was clean.ruby /tmp/with_creds.rb bundle exec rspec spec/ruby_llm/protocols/responses_gpt6_mantle_spec.rb -e 'openai.gpt-6.2'withLANG=C.UTF-8 LC_ALL=C.UTF-8andCIunset. That gave 14 of 16 passing. The two failures were the tool-loop examples, which are now gated off.CI=1:responses_gpt6_mantle_spec.rbgave 42 examples and 0 failures on seeds 1, 42, 9001 and 7.Type of change
Scope check
Required for new features
No linked issue was found. This work is tracked internally as "NewModels W1-01".
Quality check
overcommit --installand all hooks passbundle exec rake vcr:record[provider_name]bundle exec rspecmodels.json,aliases.json)AI-generated code
API changes
This is a registry and data change plus tests. It adds no new public methods or classes and changes no method signatures.