buildConfirmed13 publishers Meta released Muse Spark Open, an 816B-parameter open-weight model that activates 42B parameters per token. Per The New Stack, a company above $20M in revenue needs a separate Meta agreement before selling it as a hosted service.
Perspective Coverage
13 publishers
- Builder
- Builder 40%
- Operator
- Operator 32%
- Investor
- Investor 28%
Reality
- Evidence55
- Adoption30
- Hype gap+35
- Incentives65
- Confidence65
buildOne report1 publisher Haiku 5.5's 100K-token cliff cuts its price advantage over Sonnet 5.5 from 20x to 4x
Anthropic's Claude Haiku 5.5 costs $0.10/$0.50 per million tokens up to 100,000 input tokens, then $0.50/$2.50 on the whole request. A dev.to breakdown advises sending Haiku only narrow, verifiable subagent work and keeping Sonnet 5.5 as the default coder.
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap+35
- Incentives
- Not scored
- Confidence60
buildOne report1 publisher Claude Sonnet 5.5 outscores Opus 5.5 on Terminal-Bench 4.0, according to a dev.to post
Anthropic's Claude Sonnet 5.5 resolves 70.6% of Terminal-Bench 4.0 tasks against 66.4% for Opus 5.5, a dev.to post reports. If that gap holds on a team's own tasks, the mid-tier model is the one to test first as the default for terminal agents.
Reality
- Evidence20
- Adoption
- Not scored
- Hype gap+45
- Incentives
- Not scored
- Confidence25
buildConfirmed22 publishers Anthropic launched Claude Haiku 5.5 at an average price about 75% below Haiku 4.5. The saving varies widely with prompt length, so teams moving classification, support or query traffic need to price their own requests before they switch models.
Perspective Coverage
22 publishers
- Builder
- Builder 47%
- Operator
- Operator 36%
- Investor
- Investor 17%
Reality
- Evidence66
- Adoption30
- Hype gap+25
- Incentives70
- Confidence65
buildConfirmed8 publishers SpaceXAI lists Grok 4.7 at $2 and $6 per million tokens against $10 and $50 for GPT-6 Astra and Claude Fable 5.1, and the only per-task figures in the record put a Cursor run at $4.69 against about $9.
Perspective Coverage
8 publishers
- Builder
- Builder 51%
- Operator
- Operator 32%
- Investor
- Investor 17%
Reality
- Evidence60
- Adoption30
- Hype gap+25
- Incentives65
- Confidence58
buildConfirmed6 publishers The quarter-size global KV cache and eighth-size persistent storage are serving-cost claims rather than benchmark scores, and banking the second one requires a prefix-cache tier your stack has to already run.
Perspective Coverage
6 publishers
- Builder
- Builder 53%
- Operator
- Operator 33%
- Investor
- Investor 14%
Reality
- Evidence66
- Adoption42
- Hype gap+14
- Incentives64
- Confidence63
buildOne report1 publisher Google gives Gemini 4 Argon to vetted cyber defenders before paid API customers
Google DeepMind's Gemini 4 Argon, priced at half of Claude Opus 5.5, goes first to vetted cyber defenders, and paid API customers have no date. Until Google sets one, teams can only check its benchmark claims on paper.
Reality
- Evidence35
- Adoption15
- Hype gap+30
- Incentives65
- Confidence35
buildOne report1 publisher Failed runs decide whether Claude Sonnet 5.5 undercuts Opus 5.5
Anthropic prices Claude Sonnet 5.5 at half Opus 5.5's per-token rate, but Artificial Analysis measured $7.67 per task at max effort against Opus's $5.98. In The New Stack's own tests, runs that were billed but never finished decided which model cost less per completed job.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+20
- Incentives35
- Confidence50
buildConfirmed4 publishers Z.ai says the base model did not change between GLM-5.2 and GLM-5.3, so the coding jump and the doubled exploitation score come out of the same post-training run. Security teams inherit the second half.
Perspective Coverage
4 publishers
- Builder
- Builder 49%
- Operator
- Operator 32%
- Investor
- Investor 19%
Reality
- Evidence50
- Adoption30
- Hype gap+30
- Incentives70
- Confidence60
buildConfirmed13 publishers Reflection AI says Beam, its 501-billion-parameter open-weight model, needs three to four times less compute than comparable open models. Buyers have only the company's figures until the weights ship later in October.
Perspective Coverage
13 publishers
- Builder
- Builder 37%
- Operator
- Operator 35%
- Investor
- Investor 28%
Reality
- Evidence35
- Adoption8
- Hype gap+40
- Incentives70
- Confidence55
buildOne report1 publisher Gemini 4 Argon's per-task bill doubles when Google's launch discount ends
Google is charging $2 and $10 per million input and output tokens for Gemini 4 Argon at launch, half the $4 and $20 it will charge later. For a model that writes long answers and is open only to vetted cyber defenders, cost per task is the figure to check before switching.
Reality
- Evidence45
- Adoption15
- Hype gap+20
- Incentives60
- Confidence40
buildOne report1 publisher Reflection's Beam splits two coding benchmarks with GLM 5.2 before its weights ship
Reflection AI's 501-billion-parameter Beam beats GLM 5.2 on SWE-bench Pro and trails it on Terminal-Bench, by the company's own scores. Its claimed compute saving covers only part of a serving bill, so planning around it as a US-built open model has to wait for outside tests.
Reality
- Evidence35
- Adoption8
- Hype gap+30
- Incentives75
- Confidence40
buildOne report1 publisher Rewriting scaffolded agent runs before fine-tuning lifts a 27B model to 74.2% on Terminal-Bench 2
IntelligenceLab and Maryland researchers say rewriting scaffolded agent runs lifted Qwen-3.8-27B to 74.2% on Terminal-Bench 2, versus 53.4% from raw traces. The result argues for replaying harness-assisted wins in a plain shell before training on them.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+20
- Incentives
- Not scored
- Confidence40
buildOne report1 publisher GLM 5.3 adds no parameters, and its sharpest jump is in finding vulnerabilities
Zhipu says every gain in GLM 5.3 came from post-training on the same ~744B MoE it shipped as GLM 5.2. The largest claimed leap, in vulnerability discovery, is also the least quantified.
Reality
- Evidence25
- Adoption10
- Hype gap+35
- Incentives70
- Confidence35
Anthropic released Sonnet 5.5 at $2 and $10 per million input and output tokens, half the Opus 5.5 rate. How much a buyer saves by moving work down a tier depends on tokens burned per task and on cache reads priced identically on both models.
Perspective Coverage
5 publishers
- Builder
- Builder 46%
- Operator
- Operator 38%
- Investor
- Investor 16%
Reality
- Evidence55
- Adoption35
- Hype gap+20
- Incentives70
- Confidence60
buildOne report1 publisher NVIDIA let an optimizer agent find the four token savers in its SoL-Pi harness for Pi
NVIDIA released SoL-Pi, an MIT-licensed Pi extension with four token-saving harness mechanisms that an optimizer agent found. A dev.to review puts the saving near one third on long sessions, bought with a few lost solves.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+20
- Incentives
- Not scored
- Confidence45
buildConfirmed14 publishers Vercel's AI Gateway now routes Claude Sonnet 5.5 through a single model ID, according to a dev.to review of the week's releases. The benchmark and cost figures come only from that third-party review, so a team's own tests decide when regulated workloads move.
Perspective Coverage
14 publishers
- Builder
- Builder 47%
- Operator
- Operator 30%
- Investor
- Investor 23%
Reality
- Evidence58
- Adoption48
- Hype gap+35
- Incentives62
- Confidence60
Google's open-source RRSI rewrites agent scaffolding to add six points on Terminal-Bench
Google Cloud AI Research has released RRSI, an Apache 2.0 tool whose self-rewriting agent harness lifted Terminal-Bench 2.1 scores from 74.2% to 80.2%. Any team can use it commercially, though on tasks the agent never trained against the reported gain falls to 4.7 points.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+15
- Incentives50
- Confidence40
buildConfirmed3 publishers Fireworks' Ember-1 used 23% fewer reasoning tokens than Kimi K3 in The New Stack's tests, yet Kimi on the cheapest host would cost $1.96 to Ember's $2.48. Ember beats Fireworks' own Kimi rate and loses at the cheapest, so buyers have to price the host before the model.
Perspective Coverage
3 publishers
- Builder
- Builder 52%
- Operator
- Operator 30%
- Investor
- Investor 18%
Reality
- Evidence55
- Adoption30
- Hype gap+25
- Incentives70
- Confidence58
buildOne report1 publisher Claude Opus 5.5 ties Fable 5.1 on two coding tests while costing 28% to 50% less per run
Claude Opus 5.5 matched Fable 5.1 on every hidden test in two New Stack coding trials, at $0.75 and $1.42 per run against $1.50 and $1.96. It needed more tokens and more minutes to get there, so the saving holds only for tasks that resemble these.
Reality
- Evidence52
- Adoption
- Not scored
- Hype gap+15
- Incentives40
- Confidence55
Earlier coverage
- Bedrock serves xAI's Grok 4.7 through the Converse, Responses and Chat Completions APIs
Build · September 28, 2026 · Confirmed2 publishers
- Sonnet 5.5 comes within two points of Opus 5.5 at half the per-token price
Product · September 30, 2026 · Confirmed11 publishers
- DeepSeek open-sources the harness, then raises the price of the model
Build · August 15, 2026 · Confirmed4 publishers
- GLM-5.3 keeps GLM-5.2's base model and claims 50% more on coding: plan for shorter eval cycles
Build · August 16, 2026 · Confirmed5 publishers
- Anthropic Cuts Cache-Read Prices by 75%; Cache Reads Were ~60% of a Heavy Agent's Bill Before the Cut
Invest · September 1, 2026 · Confirmed2 publishers
- Gemini 3.8 Flash's introductory price doubles on December 31, 2026
Build · September 2, 2026 · Confirmed8 publishers
- Spark 1.3's index jump lands on the three tests that carry half the score
Build · September 3, 2026 · Confirmed6 publishers
- xAI holds Grok's $2 token price for a model 40 Elo points behind Fable 5.1
Invest · September 21, 2026 · Confirmed3 publishers
- Opus 5.5 diverts most cybersecurity requests to the older Opus 4.8
Leadership · September 22, 2026 · Confirmed2 publishers
- Price cuts minutes apart send agent routing back to the spreadsheet
Product · September 22, 2026 · Confirmed5 publishers
- Box measured Claude Opus 5.5 using a third of the tokens Opus 5 needed
Product · September 22, 2026 · Confirmed7 publishers
- Anthropic prices its newer Sonnet a third below Sonnet 4.5
Build · September 20, 2026 · One report1 publisher