Skip to content

fix(cli): price GPT Image 2.6 and every GPT Image reference picture - #7307

Merged
jrusso1020 merged 2 commits into
mainfrom
fix/cli-gpt-image-2-6-estimates
Nov 6, 2026
Merged

jrusso1020 merged 2 commits into
mainfrom
fix/cli-gpt-image-2-6-estimates

Conversation

@jrusso1020

@jrusso1020 jrusso1020 commented Nov 6, 2026 •

Copy link
Copy Markdown
Collaborator

What

media-use generate --type image now prices GPT Image 2.6 and 2.6-flare, and counts every reference picture on a GPT Image edit at about $0.0125. Before, it charged $0.0082 a reference after the first. nano-banana-2.1 gets its own price. A model the CLI doesn't know is reported as "not priced ahead" instead of being priced as GPT Image 2.5. The media-use skill's price table matches.

Why

  • References: fal bills every input picture as tokens, the first one too. Prod 2.5 edits at 16:9 cost about $0.0525 with one reference and $0.0646 with two, against $0.0402 with none (30 days to 2026-11-06). The HyperFrames price book now charges 12.5 credits for every reference (experiment-framework#65799).
  • 2.6: experiment-framework#65799 added gpt-image-2.6 and gpt-image-2.6-flare to /v3/images at 2.5's prices. 2.6 bills the same tokens as 2.5.
  • Other models: everything that wasn't nano-banana-2 was priced as 2.5. So nano-banana-2.1 (33.6 credits plus 1.68 a reference) was quoted at GPT rates, and --max-credits checked against the wrong number.

Example: a 16:9 GPT Image edit with one reference was quoted at 40 credits, the same as no reference. The price book now charges 53.

Related work

  • experiment-framework#65799 (merged): the 2.6 models and the reference rows.
  • experiment-framework#65869 (open): the HF pool charges what fal reports, so these numbers stay estimates.
  • The hyperframes-internal creation-pricing PR makes the same change.

How

  • pricing.ts: explicit image models. GPT Image is the per-aspect output price plus $0.0125 times the number of references. nano-banana-2.1 is $0.0336 plus $0.00168 a reference. An unknown model returns null, so --max-credits refuses it rather than guessing.
  • --model help and heygen.md: list the five image models.

Test plan

  • Unit tests added/updated: each GPT Image model at 16:9 costs 40 credits with no references, 53 with one and 240 with 16. nano-banana-2.1 with two references costs 37. An unknown model has null credits. The existing 2.5 case moves from $0.07 to $0.09.
  • src/heygen/generate.test.ts passes, 55 tests.
  • Static checks: oxfmt, oxlint, tsc (no errors in touched files), fallow, lint-skills, check-skill-mirror and check:studio-copy-dashes pass.
  • Documentation updated: skills/media-use/references/heygen.md.
  • Comments say where the numbers come from.

— Rames

🤖 Generated with Claude Code

The media-use image estimate priced every non-default model as GPT Image 2.5 and
charged references at $0.0082 after the first. fal bills every reference picture,
the first one too, at about $0.0125, and the price book now does the same. GPT
Image 2.6 and 2.6-flare take 2.5's prices; nano-banana-2.1 gets its own; any
other model is not priced ahead instead of guessed.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

@heygengenesis heygengenesis Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Verdict: COMMENT (needs fix) · 0 blocker · 1 important · 0 nits
The numbers match image_catalog.py and the hf_image price-book rows on EF master (12.5 per reference, 33.6 + 1.68 for nano-banana-2.1, one round-up). The Comments ratchet is red because of this diff (inline). It isn't a required check, but I'd approve once it's green and Windows render finishes.

Scope

Read the full diff and pricing.ts at the head. Checked against EF image_catalog.py (_GPT_VENDOR_COST, nano-banana-2.1 VendorCostEstimate) and the price book's hf_image rules: -reference-v2 rows for 2.5, plus 2.6/flare output and reference rows. Recomputed the tests' 40/53/240/37 credits and the $0.09 case. No other 0.0082 or reference-pricing site in the repo. CI at this head: Test (cli 1/2, 2/2), Typecheck, Build and Skills manifest are green. Comments is red. Flake gate and Windows render were still running.

— Genesis

Comment thread packages/cli/src/heygen/pricing.ts Resolved
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

@heygengenesis heygengenesis Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Verdict: APPROVE — 1 resolved, 0 open, 0 new · 7f807bd → 43ae6fa

@jrusso1020
jrusso1020 added this pull request to the merge queue Nov 6, 2026
@jrusso1020
jrusso1020 removed this pull request from the merge queue due to a manual request Nov 6, 2026
@jrusso1020
jrusso1020 added this pull request to the merge queue Nov 6, 2026
Merged via the queue into main with commit ab50e9c Nov 6, 2026
70 checks passed
@jrusso1020
jrusso1020 deleted the fix/cli-gpt-image-2-6-estimates branch November 6, 2026 22:17
SanHsien added a commit to SanHsien/hyperframes that referenced this pull request Nov 7, 2026
…-com#7336-heygen-com#7358

Adopt heygen-com#7304 (without CLI renderDiagnostic, needs heygen-com#6725), heygen-com#7331, heygen-com#7332, heygen-com#7334,
heygen-com#7337, heygen-com#7336, heygen-com#7339, heygen-com#7289, heygen-com#5594, heygen-com#6900, heygen-com#7322, heygen-com#7342, heygen-com#7343, heygen-com#7350 (onProgress
only), heygen-com#7352, heygen-com#7348, heygen-com#7344, heygen-com#7353, heygen-com#7356 and heygen-com#7355. Defer heygen-com#5708 (needs heygen-com#6911) and
heygen-com#7345 (READ_VERSION 4 vs 7); heygen-com#7307 not applicable. Re-read 14 moved pins, pin 6
new PRs (monitored 126). Make checkReportFile module-private and suppress the
cross-package closeTestChrome duplicate for fallow.

Co-authored-by: Miguel Ángel <miguel.sierra@heygen.com>
Co-authored-by: SophiaHuangCc <129631224+SophiaHuangCc@users.noreply.github.com>
Co-authored-by: Dante <duanjl.china@gmail.com>
Co-authored-by: Muzzy5150 <muzzy5150yt@gmail.com>
Co-authored-by: ump45nose <52391318+ump45nose@users.noreply.github.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant